DATA 621 HW4
Sarah Wigodsky
November 21, 2018
Multiple Logistic Regression and Binary Logistic Regression
DATA EXPLORATION and DATA PREPARATION
The data set contains information of 8,161 customers at an auto insurance company. The data set includes variables to create a model to predict the likelihood a customer will get into a car accident, TARGET_FLAG. TARGET_FLAG is a binary response variable, in which 1 represents a person who gets into a car crash and 0 represents a person who does not get into a car crash. In addition, a model will be built that will predict the cost of a claim, TARGET_AMT. The TARGET_AMT is zero if a person does not get into a crash and varies depending on the severity of the accident.
## INDEX TARGET_FLAG TARGET_AMT KIDSDRIV AGE HOMEKIDS YOJ INCOME PARENT1
## 1 1 0 0 0 60 0 11 $67,349 No
## 2 2 0 0 0 43 0 11 $91,449 No
## 3 4 0 0 0 35 1 10 $16,039 No
## 4 5 0 0 0 51 0 14 No
## 5 6 0 0 0 50 0 NA $114,986 No
## 6 7 1 2946 0 34 1 12 $125,301 Yes
## HOME_VAL MSTATUS SEX EDUCATION JOB TRAVTIME CAR_USE
## 1 $0 z_No M PhD Professional 14 Private
## 2 $257,252 z_No M z_High School z_Blue Collar 22 Commercial
## 3 $124,191 Yes z_F z_High School Clerical 5 Private
## 4 $306,251 Yes M <High School z_Blue Collar 32 Private
## 5 $243,925 Yes z_F PhD Doctor 36 Private
## 6 $0 z_No z_F Bachelors z_Blue Collar 46 Commercial
## BLUEBOOK TIF CAR_TYPE RED_CAR OLDCLAIM CLM_FREQ REVOKED MVR_PTS
## 1 $14,230 11 Minivan yes $4,461 2 No 3
## 2 $14,940 1 Minivan yes $0 0 No 0
## 3 $4,010 4 z_SUV no $38,690 2 No 3
## 4 $15,440 7 Minivan yes $0 0 No 0
## 5 $18,000 1 z_SUV no $19,217 2 Yes 3
## 6 $17,430 1 Sports Car no $0 0 No 0
## CAR_AGE URBANICITY
## 1 18 Highly Urban/ Urban
## 2 1 Highly Urban/ Urban
## 3 10 Highly Urban/ Urban
## 4 6 Highly Urban/ Urban
## 5 17 Highly Urban/ Urban
## 6 7 Highly Urban/ UrbanThe variables that will be used to build the model are:
- AGE: Age of Driver
- BLUEBOOK: Value of Vehicle
- CAR_AGE: Vehicle Age
- CAR_TYPE: Type of Car (Minivan, Panel Truck, SUV, Van, Pickup, Sports Car)
- CAR_USE: Vehicle Use (Commercial or Private)
- CLM_FREQ: Number of Claims in the Past 5 Years
- EDUCATION: Maximum Education Level (<High School, High School, Bachelors, Masters, PhD)
- HOMEKIDS: Number of Children at Home
- HOME_VAL: Home Value
- INCOME: Income
- JOB: Job Category (Professional, Blue Collar, Doctor, Clerical, Home Maker, Lawyer, Manager, Student)
- KIDSDRIV: Number of Driving Children
- MSTATUS: Marital Status
- MVR_PTS: Motor Vehicle Record Points
- OLDCLAIM: Total Payout for Claims in the Past 5 Years
- PARENT1: Single Parent
- RED_CAR: A Red Car
- REVOKED: License Revoked (Past 7 Years)
- SEX: Gender
- TIF: Number of Years the Person has Been a Customer (Time in Force)
- TRAVTIME: Distance to Work
- URBANICITY: Home/Work Area (Highly Urban/Urban, Highly Rural/Rural)
- YOJ: Years on the Job
The variables of INCOME, HOME_VAL, BLUEBOOK, and OLDCLAIM are converted from characters into numeric values to explore the data.
PARENT1 is converted into a dummy variable in which being a single parent is designated as 1 and not being a single parent is designated as 0.
MSTATUS is converted into a dummy variable in which being single is designated as 0 and being married is designated as 1.
SEX is converted into a dummy variable in which male is designated as 0 and female is designated as 1.
CARUSE is converted into a dummy variable in which private is designated as 0 and commerical is designated as 1.
RED_CAR is converted into a dummy variable in which no is designated as 0 and yes is designated as 1.
URBANICITY is converted into a dummy variable in which rural is designated as 0 and urban is designated as 1.
EDUCATION is converted into numeric variables in which 
Income
The data for income is skewed to the right, with outliers being customers who make above about $86,000. The mean is therefore higher than the median. There are 464 missing values for income. I will impute the median income for missing income values.
Year On The Job
There are a high number of people who have 0 years on the job. These people are unemployed, have been working for 12 months or less or are students. People who have been on the job for more than about 19 years and fewer than about 3 years are outliers. There are 454 missing values. I will impute the median value for the missing values.
Age
The distribution for age looks normal. There are 6 missing values. I will impute the median for the missing values.
Home Value
There are a high number of customers with home values of 0. Either that refers to renters, customers without a permanent home or there is an error in the data collection. Even if it refers to renters, the value of zero obfuscates any differentiation in the value of the home. I will remove the 0 home values and make them NAs. There are now 2,758 missing home values. About 1/3 of customers’ home values are missing.I will impute the median value for the missing values.
## Min. 1st Qu. Median Mean 3rd Qu. Max. NA's
## 50223 153074 206692 220621 270023 885282 2758
Correlation of Variables
The following are the correlation values between each of the variables. The closer the correlation is to 1 or -1, the more highly correlated the variables.
## corrplot 0.84 loaded
The variables KIDSDRIV and HOMEKIDS are positively correlated, indicating that customers with more children have more children driving.
The variables AGE and HOMEKIDS are negatively correlated, indicating that customers who are older have fewer children at home.
The variables AGE and PARENT1 are negatively correlated, indicating that customers who are single parents tend to be younger.
The variables HOMEKIDS and PARENT1 are positively correlated, indicating that customers who are single parents have more kids at home. That is a tautology.
The variables PARENT1 and MSTATUS are negatively correlated, indicating that single parents are more likely not married.
The variables YOJ and Home Maker have a positive correlation, indicating that home makers are on the job more years.
The variables INCOME and HOME_VAL are positively correlated, which makes sense. Customers with higher incomes have homes with higher values.
The variables INCOME and EDUCATION are positively correlated. Customers with higher education have higher incomes.
The variables HOME_VAL and EDUCATION are positively correlated. Customers with higher education have homes of higher values.
The variables EDUCATION and CAR_AGE are positively correlated, indicating that customers with higher education have older cars.
The variables SEX and RED_CAR are negatively correlated, indicating that men are much more likely to have red cars.
The variables SEX and z_SUV are psotively correlated, indicating that women are much more likely to have SUVs.
The variables OLDCLAIM and CLM_FREQ are positively correlated, indicating that the total number of claims in the past 5 years and the total number of claims beyond 5 years are related.
Build Models
Creating a Test Set and Training Set
Backward Elimination - Logistic Regression Model 1
## Warning: glm.fit: algorithm did not converge## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred##
## Call:
## glm(formula = train1$TARGET_FLAG ~ KIDSDRIV + YOJ + MSTATUS +
## TRAVTIME + CAR_USE + BLUEBOOK + TIF + REVOKED + MVR_PTS +
## CAR_AGE + URBANICITY + Clerical + Doctor + `Home Maker` +
## Lawyer + Manager + Professional + `z_Blue Collar` + `Panel Truck` +
## Pickup + `Sports Car` + Van + homekids_age + income_homeval_educ +
## sex_redcar_suv + oldclaim_freq, family = binomial(link = "logit"),
## data = train1)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2554 -0.7371 -0.4125 0.6777 3.0696
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.148e+00 2.541e-01 -12.392 < 2e-16 ***
## KIDSDRIV 3.204e-01 8.082e-02 3.965 7.35e-05 ***
## YOJ -2.797e-02 1.065e-02 -2.626 0.00865 **
## MSTATUS -7.363e-01 7.672e-02 -9.597 < 2e-16 ***
## TRAVTIME 1.605e-02 2.399e-03 6.692 2.21e-11 ***
## CAR_USE 8.121e-01 1.130e-01 7.185 6.70e-13 ***
## BLUEBOOK -3.528e-05 5.990e-06 -5.891 3.85e-09 ***
## TIF -5.435e-02 9.342e-03 -5.818 5.95e-09 ***
## REVOKED 7.196e-01 1.139e-01 6.317 2.66e-10 ***
## MVR_PTS 1.329e-01 1.724e-02 7.708 1.28e-14 ***
## CAR_AGE -5.745e-03 8.584e-03 -0.669 0.50332
## URBANICITY 2.425e+00 1.420e-01 17.082 < 2e-16 ***
## Clerical 1.514e-01 1.462e-01 1.036 0.30028
## Doctor -5.988e-01 3.158e-01 -1.896 0.05792 .
## `Home Maker` -5.219e-02 1.727e-01 -0.302 0.76256
## Lawyer -8.559e-02 1.754e-01 -0.488 0.62552
## Manager -1.024e+00 1.664e-01 -6.154 7.58e-10 ***
## Professional -3.396e-01 1.452e-01 -2.340 0.01929 *
## `z_Blue Collar` -1.028e-01 1.331e-01 -0.772 0.43997
## `Panel Truck` 6.121e-01 1.914e-01 3.198 0.00139 **
## Pickup 2.473e-01 1.186e-01 2.086 0.03701 *
## `Sports Car` 6.824e-01 1.274e-01 5.358 8.39e-08 ***
## Van 4.815e-01 1.521e-01 3.165 0.00155 **
## homekids_age 3.352e-03 9.966e-04 3.363 0.00077 ***
## income_homeval_educ -4.656e-07 1.162e-07 -4.007 6.14e-05 ***
## sex_redcar_suv 3.658e-01 6.747e-02 5.421 5.93e-08 ***
## oldclaim_freq 3.069e-06 4.344e-06 0.706 0.47991
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4458.7 on 4870 degrees of freedom
## AIC: 4512.7
##
## Number of Fisher Scoring iterations: 5Home Maker has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ KIDSDRIV + YOJ + MSTATUS +
## TRAVTIME + CAR_USE + BLUEBOOK + TIF + REVOKED + MVR_PTS +
## CAR_AGE + URBANICITY + Clerical + Doctor + Lawyer + Manager +
## Professional + `z_Blue Collar` + `Panel Truck` + Pickup +
## `Sports Car` + Van + homekids_age + income_homeval_educ +
## sex_redcar_suv + oldclaim_freq, family = binomial(link = "logit"),
## data = train1)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2454 -0.7362 -0.4122 0.6737 3.0709
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.172e+00 2.416e-01 -13.132 < 2e-16 ***
## KIDSDRIV 3.199e-01 8.078e-02 3.961 7.48e-05 ***
## YOJ -2.743e-02 1.050e-02 -2.613 0.008976 **
## MSTATUS -7.370e-01 7.668e-02 -9.612 < 2e-16 ***
## TRAVTIME 1.604e-02 2.399e-03 6.688 2.26e-11 ***
## CAR_USE 8.195e-01 1.104e-01 7.424 1.14e-13 ***
## BLUEBOOK -3.536e-05 5.984e-06 -5.909 3.45e-09 ***
## TIF -5.430e-02 9.339e-03 -5.814 6.10e-09 ***
## REVOKED 7.206e-01 1.139e-01 6.329 2.47e-10 ***
## MVR_PTS 1.330e-01 1.724e-02 7.718 1.18e-14 ***
## CAR_AGE -5.883e-03 8.570e-03 -0.686 0.492435
## URBANICITY 2.426e+00 1.419e-01 17.103 < 2e-16 ***
## Clerical 1.693e-01 1.339e-01 1.265 0.206018
## Doctor -5.837e-01 3.118e-01 -1.872 0.061218 .
## Lawyer -6.836e-02 1.659e-01 -0.412 0.680306
## Manager -1.009e+00 1.589e-01 -6.350 2.15e-10 ***
## Professional -3.241e-01 1.358e-01 -2.386 0.017032 *
## `z_Blue Collar` -8.968e-02 1.259e-01 -0.712 0.476181
## `Panel Truck` 6.132e-01 1.914e-01 3.204 0.001356 **
## Pickup 2.473e-01 1.186e-01 2.085 0.037045 *
## `Sports Car` 6.794e-01 1.270e-01 5.351 8.75e-08 ***
## Van 4.814e-01 1.521e-01 3.164 0.001555 **
## homekids_age 3.366e-03 9.952e-04 3.383 0.000718 ***
## income_homeval_educ -4.611e-07 1.152e-07 -4.004 6.23e-05 ***
## sex_redcar_suv 3.641e-01 6.727e-02 5.413 6.20e-08 ***
## oldclaim_freq 3.061e-06 4.345e-06 0.704 0.481217
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4458.8 on 4871 degrees of freedom
## AIC: 4510.8
##
## Number of Fisher Scoring iterations: 5Lawyer has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ KIDSDRIV + YOJ + MSTATUS +
## TRAVTIME + CAR_USE + BLUEBOOK + TIF + REVOKED + MVR_PTS +
## CAR_AGE + URBANICITY + Clerical + Doctor + Manager + Professional +
## `z_Blue Collar` + `Panel Truck` + Pickup + `Sports Car` +
## Van + homekids_age + income_homeval_educ + sex_redcar_suv +
## oldclaim_freq, family = binomial(link = "logit"), data = train1)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2452 -0.7364 -0.4122 0.6742 3.0734
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.179e+00 2.410e-01 -13.192 < 2e-16 ***
## KIDSDRIV 3.199e-01 8.079e-02 3.959 7.52e-05 ***
## YOJ -2.827e-02 1.030e-02 -2.744 0.006064 **
## MSTATUS -7.359e-01 7.662e-02 -9.604 < 2e-16 ***
## TRAVTIME 1.605e-02 2.398e-03 6.693 2.19e-11 ***
## CAR_USE 8.320e-01 1.061e-01 7.839 4.55e-15 ***
## BLUEBOOK -3.547e-05 5.980e-06 -5.931 3.01e-09 ***
## TIF -5.424e-02 9.338e-03 -5.809 6.29e-09 ***
## REVOKED 7.188e-01 1.138e-01 6.318 2.65e-10 ***
## MVR_PTS 1.329e-01 1.723e-02 7.712 1.24e-14 ***
## CAR_AGE -6.373e-03 8.489e-03 -0.751 0.452774
## URBANICITY 2.425e+00 1.419e-01 17.096 < 2e-16 ***
## Clerical 1.849e-01 1.285e-01 1.438 0.150317
## Doctor -5.455e-01 2.979e-01 -1.831 0.067047 .
## Manager -9.863e-01 1.492e-01 -6.612 3.78e-11 ***
## Professional -3.037e-01 1.265e-01 -2.399 0.016419 *
## `z_Blue Collar` -7.926e-02 1.234e-01 -0.642 0.520709
## `Panel Truck` 6.227e-01 1.900e-01 3.278 0.001047 **
## Pickup 2.484e-01 1.186e-01 2.095 0.036158 *
## `Sports Car` 6.813e-01 1.269e-01 5.369 7.90e-08 ***
## Van 4.837e-01 1.520e-01 3.182 0.001461 **
## homekids_age 3.402e-03 9.915e-04 3.431 0.000601 ***
## income_homeval_educ -4.715e-07 1.126e-07 -4.186 2.83e-05 ***
## sex_redcar_suv 3.656e-01 6.717e-02 5.443 5.24e-08 ***
## oldclaim_freq 3.123e-06 4.343e-06 0.719 0.472152
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4459.0 on 4872 degrees of freedom
## AIC: 4509
##
## Number of Fisher Scoring iterations: 5Blue Collar has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ KIDSDRIV + YOJ + MSTATUS +
## TRAVTIME + CAR_USE + BLUEBOOK + TIF + REVOKED + MVR_PTS +
## CAR_AGE + URBANICITY + Clerical + Doctor + Manager + Professional +
## `Panel Truck` + Pickup + `Sports Car` + Van + homekids_age +
## income_homeval_educ + sex_redcar_suv + oldclaim_freq, family = binomial(link = "logit"),
## data = train1)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2389 -0.7363 -0.4122 0.6803 3.0733
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.196e+00 2.397e-01 -13.333 < 2e-16 ***
## KIDSDRIV 3.166e-01 8.062e-02 3.928 8.58e-05 ***
## YOJ -3.057e-02 9.658e-03 -3.166 0.001547 **
## MSTATUS -7.337e-01 7.655e-02 -9.585 < 2e-16 ***
## TRAVTIME 1.602e-02 2.399e-03 6.681 2.38e-11 ***
## CAR_USE 8.027e-01 9.571e-02 8.387 < 2e-16 ***
## BLUEBOOK -3.574e-05 5.966e-06 -5.991 2.09e-09 ***
## TIF -5.395e-02 9.323e-03 -5.786 7.19e-09 ***
## REVOKED 7.194e-01 1.138e-01 6.324 2.55e-10 ***
## MVR_PTS 1.330e-01 1.724e-02 7.717 1.19e-14 ***
## CAR_AGE -5.548e-03 8.390e-03 -0.661 0.508424
## URBANICITY 2.421e+00 1.416e-01 17.093 < 2e-16 ***
## Clerical 2.212e-01 1.155e-01 1.916 0.055424 .
## Doctor -5.453e-01 2.979e-01 -1.831 0.067151 .
## Manager -9.626e-01 1.445e-01 -6.661 2.73e-11 ***
## Professional -2.755e-01 1.187e-01 -2.320 0.020316 *
## `Panel Truck` 6.515e-01 1.846e-01 3.529 0.000418 ***
## Pickup 2.613e-01 1.168e-01 2.237 0.025260 *
## `Sports Car` 6.841e-01 1.268e-01 5.395 6.85e-08 ***
## Van 4.961e-01 1.507e-01 3.292 0.000996 ***
## homekids_age 3.453e-03 9.879e-04 3.496 0.000473 ***
## income_homeval_educ -4.511e-07 1.078e-07 -4.184 2.87e-05 ***
## sex_redcar_suv 3.676e-01 6.711e-02 5.478 4.30e-08 ***
## oldclaim_freq 3.136e-06 4.343e-06 0.722 0.470223
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4459.4 on 4873 degrees of freedom
## AIC: 4507.4
##
## Number of Fisher Scoring iterations: 5Car age has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ KIDSDRIV + YOJ + MSTATUS +
## TRAVTIME + CAR_USE + BLUEBOOK + TIF + REVOKED + MVR_PTS +
## URBANICITY + Clerical + Doctor + Manager + Professional +
## `Panel Truck` + Pickup + `Sports Car` + Van + homekids_age +
## income_homeval_educ + sex_redcar_suv + oldclaim_freq, family = binomial(link = "logit"),
## data = train1)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2289 -0.7365 -0.4116 0.6801 3.0704
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.226e+00 2.356e-01 -13.692 < 2e-16 ***
## KIDSDRIV 3.167e-01 8.061e-02 3.929 8.52e-05 ***
## YOJ -3.051e-02 9.655e-03 -3.160 0.001577 **
## MSTATUS -7.320e-01 7.649e-02 -9.570 < 2e-16 ***
## TRAVTIME 1.602e-02 2.398e-03 6.681 2.37e-11 ***
## CAR_USE 8.094e-01 9.519e-02 8.503 < 2e-16 ***
## BLUEBOOK -3.576e-05 5.969e-06 -5.991 2.09e-09 ***
## TIF -5.402e-02 9.322e-03 -5.795 6.82e-09 ***
## REVOKED 7.199e-01 1.138e-01 6.328 2.49e-10 ***
## MVR_PTS 1.332e-01 1.723e-02 7.730 1.07e-14 ***
## URBANICITY 2.419e+00 1.416e-01 17.085 < 2e-16 ***
## Clerical 2.297e-01 1.148e-01 2.001 0.045362 *
## Doctor -5.368e-01 2.977e-01 -1.803 0.071393 .
## Manager -9.636e-01 1.446e-01 -6.666 2.63e-11 ***
## Professional -2.727e-01 1.186e-01 -2.299 0.021519 *
## `Panel Truck` 6.491e-01 1.846e-01 3.516 0.000439 ***
## Pickup 2.604e-01 1.168e-01 2.230 0.025776 *
## `Sports Car` 6.851e-01 1.268e-01 5.404 6.51e-08 ***
## Van 4.964e-01 1.507e-01 3.294 0.000989 ***
## homekids_age 3.471e-03 9.873e-04 3.516 0.000438 ***
## income_homeval_educ -4.864e-07 9.442e-08 -5.151 2.59e-07 ***
## sex_redcar_suv 3.668e-01 6.709e-02 5.468 4.56e-08 ***
## oldclaim_freq 3.073e-06 4.343e-06 0.708 0.479202
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4459.8 on 4874 degrees of freedom
## AIC: 4505.8
##
## Number of Fisher Scoring iterations: 5Old Claim/Claim Frequency has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ KIDSDRIV + YOJ + MSTATUS +
## TRAVTIME + CAR_USE + BLUEBOOK + TIF + REVOKED + MVR_PTS +
## URBANICITY + Clerical + Doctor + Manager + Professional +
## `Panel Truck` + Pickup + `Sports Car` + Van + homekids_age +
## income_homeval_educ + sex_redcar_suv, family = binomial(link = "logit"),
## data = train1)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2332 -0.7375 -0.4125 0.6741 3.0716
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.227e+00 2.356e-01 -13.694 < 2e-16 ***
## KIDSDRIV 3.164e-01 8.065e-02 3.923 8.73e-05 ***
## YOJ -3.029e-02 9.652e-03 -3.138 0.001700 **
## MSTATUS -7.333e-01 7.647e-02 -9.589 < 2e-16 ***
## TRAVTIME 1.600e-02 2.398e-03 6.672 2.52e-11 ***
## CAR_USE 8.104e-01 9.517e-02 8.515 < 2e-16 ***
## BLUEBOOK -3.577e-05 5.969e-06 -5.992 2.07e-09 ***
## TIF -5.391e-02 9.320e-03 -5.784 7.31e-09 ***
## REVOKED 7.540e-01 1.030e-01 7.324 2.41e-13 ***
## MVR_PTS 1.363e-01 1.667e-02 8.176 2.93e-16 ***
## URBANICITY 2.425e+00 1.413e-01 17.159 < 2e-16 ***
## Clerical 2.292e-01 1.148e-01 1.997 0.045844 *
## Doctor -5.380e-01 2.979e-01 -1.806 0.070915 .
## Manager -9.632e-01 1.445e-01 -6.664 2.66e-11 ***
## Professional -2.744e-01 1.186e-01 -2.314 0.020660 *
## `Panel Truck` 6.500e-01 1.846e-01 3.520 0.000431 ***
## Pickup 2.596e-01 1.168e-01 2.223 0.026187 *
## `Sports Car` 6.880e-01 1.267e-01 5.432 5.57e-08 ***
## Van 4.966e-01 1.507e-01 3.295 0.000985 ***
## homekids_age 3.458e-03 9.871e-04 3.503 0.000460 ***
## income_homeval_educ -4.885e-07 9.437e-08 -5.176 2.26e-07 ***
## sex_redcar_suv 3.661e-01 6.707e-02 5.458 4.82e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4460.3 on 4875 degrees of freedom
## AIC: 4504.3
##
## Number of Fisher Scoring iterations: 5Doctor has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ KIDSDRIV + YOJ + MSTATUS +
## TRAVTIME + CAR_USE + BLUEBOOK + TIF + REVOKED + MVR_PTS +
## URBANICITY + Clerical + Manager + Professional + `Panel Truck` +
## Pickup + `Sports Car` + Van + homekids_age + income_homeval_educ +
## sex_redcar_suv, family = binomial(link = "logit"), data = train1)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2358 -0.7370 -0.4115 0.6749 3.0747
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.232e+00 2.357e-01 -13.714 < 2e-16 ***
## KIDSDRIV 3.179e-01 8.068e-02 3.940 8.16e-05 ***
## YOJ -3.040e-02 9.647e-03 -3.151 0.001628 **
## MSTATUS -7.289e-01 7.641e-02 -9.540 < 2e-16 ***
## TRAVTIME 1.601e-02 2.398e-03 6.676 2.45e-11 ***
## CAR_USE 8.337e-01 9.448e-02 8.824 < 2e-16 ***
## BLUEBOOK -3.578e-05 5.963e-06 -6.000 1.97e-09 ***
## TIF -5.358e-02 9.321e-03 -5.749 9.00e-09 ***
## REVOKED 7.562e-01 1.029e-01 7.350 1.99e-13 ***
## MVR_PTS 1.363e-01 1.667e-02 8.174 2.97e-16 ***
## URBANICITY 2.425e+00 1.414e-01 17.150 < 2e-16 ***
## Clerical 2.347e-01 1.148e-01 2.044 0.040988 *
## Manager -9.283e-01 1.435e-01 -6.468 9.91e-11 ***
## Professional -2.485e-01 1.179e-01 -2.107 0.035086 *
## `Panel Truck` 6.677e-01 1.845e-01 3.620 0.000295 ***
## Pickup 2.567e-01 1.167e-01 2.199 0.027897 *
## `Sports Car` 6.874e-01 1.265e-01 5.435 5.49e-08 ***
## Van 4.967e-01 1.505e-01 3.300 0.000968 ***
## homekids_age 3.498e-03 9.869e-04 3.545 0.000393 ***
## income_homeval_educ -5.452e-07 8.981e-08 -6.070 1.28e-09 ***
## sex_redcar_suv 3.659e-01 6.704e-02 5.459 4.80e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4463.9 on 4876 degrees of freedom
## AIC: 4505.9
##
## Number of Fisher Scoring iterations: 5## (Intercept) KIDSDRIV YOJ
## -4.808017e-01 4.728775e-02 -4.522150e-03
## MSTATUS TRAVTIME CAR_USE
## -1.084377e-01 2.381669e-03 1.240251e-01
## BLUEBOOK TIF REVOKED
## -5.323281e-06 -7.971547e-03 1.125049e-01
## MVR_PTS URBANICITY Clerical
## 2.027693e-02 3.608180e-01 3.491427e-02
## Manager Professional `Panel Truck`
## -1.380964e-01 -3.697466e-02 9.932959e-02
## Pickup `Sports Car` Van
## 3.818790e-02 1.022702e-01 7.388903e-02
## homekids_age income_homeval_educ sex_redcar_suv
## 5.203981e-04 -8.110785e-08 5.444163e-02The variables that have an effect on whether the car was in a crash are: KIDSDRIV, YOJ, MSTATUS, TRAVTIME, CAR_USE, BLUEBOOK, TIF, REVOKED, MVR_PTS, URBANICITY, Clerical, Manager, Professional, Panel Truck, Pickup, Sports Car, Van, homekids_age =, income_homeval_edu, and sex_redcar_suv.
The marginal effects reflect the change in the probability the target equals 1 given a 1 unit change in the independent variable. The marginal effect is determined at the mean for each of the independent variables. A 1 unit increase in KIDSDRIV, the number of driving children, results in a 4.7% increase in the probability that the customer has a claim. A 1 unit increase in YOJ, years on the job, results in a .5% decrease in the probability that the customer has a claim. Being married results in a 11% decrease in the likelihood a customer has a claim. A 1 unit incrase in travel time results in a 0.2% increase in the likelihood that the customer has a claim. A customer who drives a commercial vehicle is 12% more likely to have a claim. A 1 unit increase in the bluebook value of the car results in a 0.0005% decrease in the likelihood that a customer will have a claim. A 1 unit increase in TIF, time in force, results in a 0.80 decrease in the likelihood the customer has a claim. Having a license revoked results in a 11% increase in the likelihood in the customer having a claim. A 1 unit increase in MVR_PTS, the number of points, results in a 2% increase in the likelihood the customer has a claim. A customer who lives in an urban area results in a 36% increase in the likelihood the customer will have a claim. Driving a sports car results in a 10% increase in the likelihood the customer will have a claim. Driving a can results in a 7% increase in the likelihood the customer will have a claim. A 1 unit increase in homekids_age results in a .05% increase in the likelihood a customer has a claim. A 1 unit increase in income_homeval_educ results in a .000008 decrease in the likelihood the customer has a claim. A 1 unit increase in sex_redcar_suv results in a 5.4% increase in the likelihood the customer has a claim.
Prediction from Model 1
## [1] 0.32## [1] 0.7909815The cutoff associated with a point the farthest distance from the ROC curve is 0.32. I will use 0.32 as the cutoff for making predictions. A value above 0.32 will be assigned a target of one and a value below 0.32 wil be assigned a value of zero. The area under the curve is 0.79.
Confusion Matrix
## true
## pred 0 1
## 0 1846 257
## 1 560 601The model predicted 1,846 0’s that were actually 0. The model predicted 257 0’s that were actually 1.
The model predicted 560 1’s that were actaully 0. The model predicted 601 1’s that were actually 1.
## accuracy error precision sensitivity specificity f1
## 1 0.7496936 0.2503064 0.4094034 0.7672485 0.7004662 0.5339118Logit Model 2
For the second model, I will not combine correlated varaibles. ####Creating a Test Set and Training Set
Backward Elimination - Logistic Regression Model 2
## Warning: glm.fit: algorithm did not converge## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred##
## Call:
## glm(formula = train2$TARGET_FLAG ~ KIDSDRIV + AGE + HOMEKIDS +
## YOJ + INCOME + PARENT1 + HOME_VAL + MSTATUS + SEX + EDUCATION +
## TRAVTIME + CAR_USE + BLUEBOOK + TIF + RED_CAR + OLDCLAIM +
## CLM_FREQ + REVOKED + MVR_PTS + CAR_AGE + URBANICITY + Clerical +
## Doctor + `Home Maker` + Lawyer + Manager + Professional +
## `z_Blue Collar` + `Panel Truck` + Pickup + `Sports Car` +
## Van + z_SUV, family = binomial(link = "logit"), data = train2)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2671 -0.7212 -0.4038 0.6187 3.1041
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.345e+00 3.606e-01 -9.277 < 2e-16 ***
## KIDSDRIV 3.274e-01 7.979e-02 4.103 4.07e-05 ***
## AGE -1.620e-03 5.189e-03 -0.312 0.754882
## HOMEKIDS 6.425e-02 4.721e-02 1.361 0.173521
## YOJ -1.878e-02 1.095e-02 -1.716 0.086156 .
## INCOME -6.258e-06 1.509e-06 -4.148 3.36e-05 ***
## PARENT1 3.933e-01 1.427e-01 2.755 0.005861 **
## HOME_VAL 9.030e-07 7.835e-07 1.153 0.249114
## MSTATUS -6.179e-01 9.444e-02 -6.543 6.04e-11 ***
## SEX 3.713e-02 1.427e-01 0.260 0.794692
## EDUCATION -1.292e-01 5.927e-02 -2.180 0.029280 *
## TRAVTIME 1.600e-02 2.428e-03 6.590 4.41e-11 ***
## CAR_USE 8.247e-01 1.158e-01 7.120 1.08e-12 ***
## BLUEBOOK -2.512e-05 6.909e-06 -3.636 0.000277 ***
## TIF -5.506e-02 9.426e-03 -5.841 5.19e-09 ***
## RED_CAR 9.476e-02 1.106e-01 0.857 0.391538
## OLDCLAIM -7.861e-06 4.929e-06 -1.595 0.110766
## CLM_FREQ 1.884e-01 3.679e-02 5.123 3.01e-07 ***
## REVOKED 8.459e-01 1.174e-01 7.202 5.92e-13 ***
## MVR_PTS 1.020e-01 1.808e-02 5.644 1.66e-08 ***
## CAR_AGE 8.909e-04 9.377e-03 0.095 0.924314
## URBANICITY 2.389e+00 1.448e-01 16.494 < 2e-16 ***
## Clerical 1.701e-01 1.497e-01 1.137 0.255659
## Doctor -4.867e-01 3.216e-01 -1.513 0.130254
## `Home Maker` -3.165e-02 1.821e-01 -0.174 0.862021
## Lawyer 5.958e-02 1.825e-01 0.327 0.744017
## Manager -9.211e-01 1.707e-01 -5.398 6.75e-08 ***
## Professional -2.233e-01 1.484e-01 -1.505 0.132399
## `z_Blue Collar` -2.132e-02 1.395e-01 -0.153 0.878550
## `Panel Truck` 6.250e-01 2.079e-01 3.007 0.002642 **
## Pickup 4.175e-01 1.276e-01 3.271 0.001072 **
## `Sports Car` 9.230e-01 1.676e-01 5.508 3.63e-08 ***
## Van 5.804e-01 1.614e-01 3.597 0.000322 ***
## z_SUV 7.123e-01 1.408e-01 5.058 4.23e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4391.8 on 4863 degrees of freedom
## AIC: 4459.8
##
## Number of Fisher Scoring iterations: 5Blue Collar has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train2$TARGET_FLAG ~ KIDSDRIV + AGE + HOMEKIDS +
## YOJ + INCOME + PARENT1 + HOME_VAL + MSTATUS + SEX + EDUCATION +
## TRAVTIME + CAR_USE + BLUEBOOK + TIF + RED_CAR + OLDCLAIM +
## CLM_FREQ + REVOKED + MVR_PTS + CAR_AGE + URBANICITY + Clerical +
## Doctor + `Home Maker` + Lawyer + Manager + Professional +
## `Panel Truck` + Pickup + `Sports Car` + Van + z_SUV, family = binomial(link = "logit"),
## data = train2)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2639 -0.7203 -0.4038 0.6194 3.1040
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.357e+00 3.515e-01 -9.551 < 2e-16 ***
## KIDSDRIV 3.267e-01 7.965e-02 4.102 4.10e-05 ***
## AGE -1.596e-03 5.187e-03 -0.308 0.758284
## HOMEKIDS 6.509e-02 4.689e-02 1.388 0.165073
## YOJ -1.919e-02 1.062e-02 -1.806 0.070950 .
## INCOME -6.312e-06 1.468e-06 -4.298 1.72e-05 ***
## PARENT1 3.926e-01 1.427e-01 2.752 0.005927 **
## HOME_VAL 9.191e-07 7.767e-07 1.183 0.236671
## MSTATUS -6.178e-01 9.444e-02 -6.542 6.08e-11 ***
## SEX 3.802e-02 1.426e-01 0.267 0.789715
## EDUCATION -1.262e-01 5.603e-02 -2.253 0.024258 *
## TRAVTIME 1.599e-02 2.427e-03 6.587 4.47e-11 ***
## CAR_USE 8.199e-01 1.113e-01 7.363 1.80e-13 ***
## BLUEBOOK -2.518e-05 6.900e-06 -3.649 0.000263 ***
## TIF -5.499e-02 9.413e-03 -5.841 5.18e-09 ***
## RED_CAR 9.504e-02 1.106e-01 0.859 0.390095
## OLDCLAIM -7.858e-06 4.930e-06 -1.594 0.110960
## CLM_FREQ 1.884e-01 3.678e-02 5.121 3.04e-07 ***
## REVOKED 8.460e-01 1.174e-01 7.203 5.89e-13 ***
## MVR_PTS 1.020e-01 1.808e-02 5.644 1.66e-08 ***
## CAR_AGE 9.415e-04 9.371e-03 0.100 0.919972
## URBANICITY 2.388e+00 1.447e-01 16.505 < 2e-16 ***
## Clerical 1.826e-01 1.254e-01 1.456 0.145355
## Doctor -4.814e-01 3.198e-01 -1.505 0.132279
## `Home Maker` -2.459e-02 1.762e-01 -0.140 0.889029
## Lawyer 6.656e-02 1.767e-01 0.377 0.706448
## Manager -9.117e-01 1.591e-01 -5.731 1.00e-08 ***
## Professional -2.130e-01 1.321e-01 -1.613 0.106828
## `Panel Truck` 6.330e-01 2.013e-01 3.144 0.001664 **
## Pickup 4.205e-01 1.261e-01 3.335 0.000853 ***
## `Sports Car` 9.226e-01 1.676e-01 5.506 3.66e-08 ***
## Van 5.838e-01 1.599e-01 3.652 0.000261 ***
## z_SUV 7.118e-01 1.408e-01 5.056 4.28e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4391.8 on 4864 degrees of freedom
## AIC: 4457.8
##
## Number of Fisher Scoring iterations: 5Home Maker has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train2$TARGET_FLAG ~ KIDSDRIV + AGE + HOMEKIDS +
## YOJ + INCOME + PARENT1 + HOME_VAL + MSTATUS + SEX + EDUCATION +
## TRAVTIME + CAR_USE + BLUEBOOK + TIF + RED_CAR + OLDCLAIM +
## CLM_FREQ + REVOKED + MVR_PTS + CAR_AGE + URBANICITY + Clerical +
## Doctor + Lawyer + Manager + Professional + `Panel Truck` +
## Pickup + `Sports Car` + Van + z_SUV, family = binomial(link = "logit"),
## data = train2)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2603 -0.7215 -0.4039 0.6192 3.1044
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.368e+00 3.438e-01 -9.795 < 2e-16 ***
## KIDSDRIV 3.268e-01 7.964e-02 4.104 4.07e-05 ***
## AGE -1.660e-03 5.166e-03 -0.321 0.747909
## HOMEKIDS 6.505e-02 4.688e-02 1.387 0.165292
## YOJ -1.882e-02 1.029e-02 -1.829 0.067416 .
## INCOME -6.297e-06 1.465e-06 -4.299 1.72e-05 ***
## PARENT1 3.925e-01 1.427e-01 2.751 0.005933 **
## HOME_VAL 9.415e-07 7.599e-07 1.239 0.215363
## MSTATUS -6.181e-01 9.441e-02 -6.547 5.87e-11 ***
## SEX 3.583e-02 1.417e-01 0.253 0.800385
## EDUCATION -1.278e-01 5.494e-02 -2.326 0.020032 *
## TRAVTIME 1.599e-02 2.427e-03 6.587 4.50e-11 ***
## CAR_USE 8.249e-01 1.053e-01 7.838 4.58e-15 ***
## BLUEBOOK -2.515e-05 6.897e-06 -3.647 0.000265 ***
## TIF -5.498e-02 9.413e-03 -5.841 5.19e-09 ***
## RED_CAR 9.535e-02 1.106e-01 0.862 0.388487
## OLDCLAIM -7.860e-06 4.930e-06 -1.594 0.110892
## CLM_FREQ 1.883e-01 3.678e-02 5.120 3.06e-07 ***
## REVOKED 8.465e-01 1.174e-01 7.210 5.58e-13 ***
## MVR_PTS 1.021e-01 1.807e-02 5.651 1.60e-08 ***
## CAR_AGE 9.587e-04 9.370e-03 0.102 0.918510
## URBANICITY 2.389e+00 1.445e-01 16.529 < 2e-16 ***
## Clerical 1.876e-01 1.201e-01 1.562 0.118357
## Doctor -4.737e-01 3.150e-01 -1.504 0.132648
## Lawyer 7.399e-02 1.685e-01 0.439 0.660586
## Manager -9.067e-01 1.551e-01 -5.845 5.06e-09 ***
## Professional -2.085e-01 1.281e-01 -1.628 0.103589
## `Panel Truck` 6.300e-01 2.001e-01 3.148 0.001646 **
## Pickup 4.195e-01 1.259e-01 3.332 0.000861 ***
## `Sports Car` 9.224e-01 1.675e-01 5.505 3.68e-08 ***
## Van 5.819e-01 1.593e-01 3.653 0.000259 ***
## z_SUV 7.121e-01 1.408e-01 5.058 4.23e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4391.8 on 4865 degrees of freedom
## AIC: 4455.8
##
## Number of Fisher Scoring iterations: 5CAR_AGE has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train2$TARGET_FLAG ~ KIDSDRIV + AGE + HOMEKIDS +
## YOJ + INCOME + PARENT1 + HOME_VAL + MSTATUS + SEX + EDUCATION +
## TRAVTIME + CAR_USE + BLUEBOOK + TIF + RED_CAR + OLDCLAIM +
## CLM_FREQ + REVOKED + MVR_PTS + URBANICITY + Clerical + Doctor +
## Lawyer + Manager + Professional + `Panel Truck` + Pickup +
## `Sports Car` + Van + z_SUV, family = binomial(link = "logit"),
## data = train2)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2610 -0.7220 -0.4036 0.6189 3.1047
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.365e+00 3.426e-01 -9.821 < 2e-16 ***
## KIDSDRIV 3.268e-01 7.963e-02 4.104 4.06e-05 ***
## AGE -1.658e-03 5.166e-03 -0.321 0.748244
## HOMEKIDS 6.506e-02 4.688e-02 1.388 0.165259
## YOJ -1.881e-02 1.029e-02 -1.828 0.067504 .
## INCOME -6.293e-06 1.464e-06 -4.298 1.72e-05 ***
## PARENT1 3.926e-01 1.427e-01 2.752 0.005931 **
## HOME_VAL 9.389e-07 7.594e-07 1.236 0.216344
## MSTATUS -6.182e-01 9.441e-02 -6.549 5.81e-11 ***
## SEX 3.603e-02 1.417e-01 0.254 0.799267
## EDUCATION -1.249e-01 4.697e-02 -2.658 0.007855 **
## TRAVTIME 1.599e-02 2.427e-03 6.586 4.51e-11 ***
## CAR_USE 8.244e-01 1.051e-01 7.844 4.36e-15 ***
## BLUEBOOK -2.515e-05 6.897e-06 -3.647 0.000265 ***
## TIF -5.497e-02 9.412e-03 -5.840 5.22e-09 ***
## RED_CAR 9.560e-02 1.105e-01 0.865 0.387111
## OLDCLAIM -7.856e-06 4.930e-06 -1.594 0.111041
## CLM_FREQ 1.884e-01 3.678e-02 5.123 3.01e-07 ***
## REVOKED 8.463e-01 1.174e-01 7.210 5.61e-13 ***
## MVR_PTS 1.021e-01 1.806e-02 5.650 1.60e-08 ***
## URBANICITY 2.389e+00 1.445e-01 16.529 < 2e-16 ***
## Clerical 1.872e-01 1.201e-01 1.559 0.119013
## Doctor -4.751e-01 3.147e-01 -1.510 0.131080
## Lawyer 7.522e-02 1.681e-01 0.448 0.654495
## Manager -9.066e-01 1.551e-01 -5.845 5.07e-09 ***
## Professional -2.087e-01 1.280e-01 -1.630 0.103066
## `Panel Truck` 6.304e-01 2.001e-01 3.151 0.001629 **
## Pickup 4.195e-01 1.259e-01 3.333 0.000860 ***
## `Sports Car` 9.221e-01 1.675e-01 5.505 3.70e-08 ***
## Van 5.818e-01 1.593e-01 3.652 0.000260 ***
## z_SUV 7.120e-01 1.408e-01 5.058 4.24e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4391.8 on 4866 degrees of freedom
## AIC: 4453.8
##
## Number of Fisher Scoring iterations: 5SEX has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train2$TARGET_FLAG ~ KIDSDRIV + AGE + HOMEKIDS +
## YOJ + INCOME + PARENT1 + HOME_VAL + MSTATUS + EDUCATION +
## TRAVTIME + CAR_USE + BLUEBOOK + TIF + RED_CAR + OLDCLAIM +
## CLM_FREQ + REVOKED + MVR_PTS + URBANICITY + Clerical + Doctor +
## Lawyer + Manager + Professional + `Panel Truck` + Pickup +
## `Sports Car` + Van + z_SUV, family = binomial(link = "logit"),
## data = train2)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2609 -0.7203 -0.4042 0.6190 3.1031
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.349e+00 3.369e-01 -9.942 < 2e-16 ***
## KIDSDRIV 3.272e-01 7.963e-02 4.109 3.98e-05 ***
## AGE -1.789e-03 5.140e-03 -0.348 0.727755
## HOMEKIDS 6.538e-02 4.687e-02 1.395 0.162992
## YOJ -1.888e-02 1.028e-02 -1.836 0.066381 .
## INCOME -6.298e-06 1.464e-06 -4.302 1.69e-05 ***
## PARENT1 3.926e-01 1.427e-01 2.752 0.005931 **
## HOME_VAL 9.333e-07 7.591e-07 1.230 0.218880
## MSTATUS -6.182e-01 9.440e-02 -6.549 5.80e-11 ***
## EDUCATION -1.243e-01 4.693e-02 -2.650 0.008060 **
## TRAVTIME 1.599e-02 2.427e-03 6.586 4.51e-11 ***
## CAR_USE 8.232e-01 1.050e-01 7.840 4.49e-15 ***
## BLUEBOOK -2.449e-05 6.387e-06 -3.835 0.000126 ***
## TIF -5.499e-02 9.412e-03 -5.843 5.14e-09 ***
## RED_CAR 8.281e-02 9.833e-02 0.842 0.399693
## OLDCLAIM -7.886e-06 4.929e-06 -1.600 0.109592
## CLM_FREQ 1.885e-01 3.677e-02 5.125 2.98e-07 ***
## REVOKED 8.462e-01 1.174e-01 7.209 5.64e-13 ***
## MVR_PTS 1.021e-01 1.806e-02 5.650 1.60e-08 ***
## URBANICITY 2.389e+00 1.445e-01 16.529 < 2e-16 ***
## Clerical 1.860e-01 1.200e-01 1.551 0.121008
## Doctor -4.795e-01 3.142e-01 -1.526 0.126988
## Lawyer 7.335e-02 1.679e-01 0.437 0.662253
## Manager -9.079e-01 1.550e-01 -5.858 4.70e-09 ***
## Professional -2.096e-01 1.280e-01 -1.637 0.101583
## `Panel Truck` 6.142e-01 1.896e-01 3.240 0.001196 **
## Pickup 4.200e-01 1.259e-01 3.337 0.000848 ***
## `Sports Car` 9.421e-01 1.480e-01 6.365 1.95e-10 ***
## Van 5.721e-01 1.546e-01 3.700 0.000216 ***
## z_SUV 7.315e-01 1.183e-01 6.182 6.32e-10 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4391.9 on 4867 degrees of freedom
## AIC: 4451.9
##
## Number of Fisher Scoring iterations: 5Age has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train2$TARGET_FLAG ~ KIDSDRIV + HOMEKIDS + YOJ +
## INCOME + PARENT1 + HOME_VAL + MSTATUS + EDUCATION + TRAVTIME +
## CAR_USE + BLUEBOOK + TIF + RED_CAR + OLDCLAIM + CLM_FREQ +
## REVOKED + MVR_PTS + URBANICITY + Clerical + Doctor + Lawyer +
## Manager + Professional + `Panel Truck` + Pickup + `Sports Car` +
## Van + z_SUV, family = binomial(link = "logit"), data = train2)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2689 -0.7197 -0.4055 0.6184 3.0972
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.420e+00 2.687e-01 -12.728 < 2e-16 ***
## KIDSDRIV 3.220e-01 7.826e-02 4.115 3.87e-05 ***
## HOMEKIDS 7.163e-02 4.327e-02 1.655 0.097836 .
## YOJ -1.941e-02 1.017e-02 -1.909 0.056238 .
## INCOME -6.258e-06 1.459e-06 -4.290 1.78e-05 ***
## PARENT1 3.984e-01 1.417e-01 2.812 0.004930 **
## HOME_VAL 9.167e-07 7.574e-07 1.210 0.226114
## MSTATUS -6.194e-01 9.434e-02 -6.566 5.16e-11 ***
## EDUCATION -1.256e-01 4.679e-02 -2.683 0.007288 **
## TRAVTIME 1.597e-02 2.427e-03 6.581 4.67e-11 ***
## CAR_USE 8.233e-01 1.050e-01 7.841 4.48e-15 ***
## BLUEBOOK -2.475e-05 6.344e-06 -3.901 9.58e-05 ***
## TIF -5.496e-02 9.411e-03 -5.840 5.23e-09 ***
## RED_CAR 8.168e-02 9.826e-02 0.831 0.405829
## OLDCLAIM -7.878e-06 4.929e-06 -1.598 0.109996
## CLM_FREQ 1.880e-01 3.676e-02 5.116 3.13e-07 ***
## REVOKED 8.458e-01 1.174e-01 7.205 5.82e-13 ***
## MVR_PTS 1.024e-01 1.804e-02 5.672 1.41e-08 ***
## URBANICITY 2.391e+00 1.444e-01 16.552 < 2e-16 ***
## Clerical 1.883e-01 1.198e-01 1.572 0.116040
## Doctor -4.845e-01 3.138e-01 -1.544 0.122621
## Lawyer 7.043e-02 1.677e-01 0.420 0.674549
## Manager -9.106e-01 1.548e-01 -5.883 4.03e-09 ***
## Professional -2.094e-01 1.280e-01 -1.636 0.101752
## `Panel Truck` 6.164e-01 1.895e-01 3.254 0.001139 **
## Pickup 4.198e-01 1.259e-01 3.335 0.000853 ***
## `Sports Car` 9.382e-01 1.476e-01 6.358 2.05e-10 ***
## Van 5.731e-01 1.546e-01 3.707 0.000210 ***
## z_SUV 7.291e-01 1.181e-01 6.173 6.68e-10 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4392.0 on 4868 degrees of freedom
## AIC: 4450
##
## Number of Fisher Scoring iterations: 5Lawyer has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train2$TARGET_FLAG ~ KIDSDRIV + HOMEKIDS + YOJ +
## INCOME + PARENT1 + HOME_VAL + MSTATUS + EDUCATION + TRAVTIME +
## CAR_USE + BLUEBOOK + TIF + RED_CAR + OLDCLAIM + CLM_FREQ +
## REVOKED + MVR_PTS + URBANICITY + Clerical + Doctor + Manager +
## Professional + `Panel Truck` + Pickup + `Sports Car` + Van +
## z_SUV, family = binomial(link = "logit"), data = train2)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2669 -0.7201 -0.4047 0.6221 3.0944
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.419e+00 2.686e-01 -12.730 < 2e-16 ***
## KIDSDRIV 3.218e-01 7.823e-02 4.113 3.91e-05 ***
## HOMEKIDS 7.050e-02 4.317e-02 1.633 0.102493
## YOJ -1.885e-02 1.008e-02 -1.871 0.061347 .
## INCOME -6.225e-06 1.456e-06 -4.277 1.90e-05 ***
## PARENT1 3.982e-01 1.417e-01 2.811 0.004936 **
## HOME_VAL 9.200e-07 7.566e-07 1.216 0.224006
## MSTATUS -6.199e-01 9.434e-02 -6.571 5.00e-11 ***
## EDUCATION -1.180e-01 4.311e-02 -2.737 0.006206 **
## TRAVTIME 1.596e-02 2.427e-03 6.575 4.87e-11 ***
## CAR_USE 8.062e-01 9.669e-02 8.338 < 2e-16 ***
## BLUEBOOK -2.468e-05 6.340e-06 -3.893 9.91e-05 ***
## TIF -5.497e-02 9.411e-03 -5.842 5.17e-09 ***
## RED_CAR 8.264e-02 9.824e-02 0.841 0.400206
## OLDCLAIM -7.930e-06 4.927e-06 -1.610 0.107486
## CLM_FREQ 1.878e-01 3.676e-02 5.110 3.23e-07 ***
## REVOKED 8.474e-01 1.173e-01 7.222 5.13e-13 ***
## MVR_PTS 1.025e-01 1.804e-02 5.681 1.34e-08 ***
## URBANICITY 2.391e+00 1.444e-01 16.557 < 2e-16 ***
## Clerical 1.785e-01 1.174e-01 1.520 0.128561
## Doctor -5.257e-01 2.980e-01 -1.764 0.077686 .
## Manager -9.317e-01 1.463e-01 -6.367 1.93e-10 ***
## Professional -2.276e-01 1.204e-01 -1.891 0.058644 .
## `Panel Truck` 6.112e-01 1.890e-01 3.234 0.001220 **
## Pickup 4.205e-01 1.258e-01 3.341 0.000834 ***
## `Sports Car` 9.363e-01 1.475e-01 6.347 2.19e-10 ***
## Van 5.723e-01 1.546e-01 3.701 0.000214 ***
## z_SUV 7.277e-01 1.181e-01 6.163 7.13e-10 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4392.2 on 4869 degrees of freedom
## AIC: 4448.2
##
## Number of Fisher Scoring iterations: 5RED_CAR has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train2$TARGET_FLAG ~ KIDSDRIV + HOMEKIDS + YOJ +
## INCOME + PARENT1 + HOME_VAL + MSTATUS + EDUCATION + TRAVTIME +
## CAR_USE + BLUEBOOK + TIF + OLDCLAIM + CLM_FREQ + REVOKED +
## MVR_PTS + URBANICITY + Clerical + Doctor + Manager + Professional +
## `Panel Truck` + Pickup + `Sports Car` + Van + z_SUV, family = binomial(link = "logit"),
## data = train2)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2663 -0.7182 -0.4054 0.6228 3.1106
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.368e+00 2.616e-01 -12.876 < 2e-16 ***
## KIDSDRIV 3.207e-01 7.817e-02 4.103 4.08e-05 ***
## HOMEKIDS 7.048e-02 4.316e-02 1.633 0.102473
## YOJ -1.853e-02 1.007e-02 -1.841 0.065693 .
## INCOME -6.256e-06 1.455e-06 -4.298 1.72e-05 ***
## PARENT1 3.954e-01 1.416e-01 2.792 0.005237 **
## HOME_VAL 9.258e-07 7.564e-07 1.224 0.220925
## MSTATUS -6.209e-01 9.433e-02 -6.582 4.64e-11 ***
## EDUCATION -1.183e-01 4.310e-02 -2.746 0.006034 **
## TRAVTIME 1.595e-02 2.426e-03 6.572 4.95e-11 ***
## CAR_USE 8.065e-01 9.668e-02 8.341 < 2e-16 ***
## BLUEBOOK -2.590e-05 6.173e-06 -4.196 2.71e-05 ***
## TIF -5.496e-02 9.406e-03 -5.843 5.12e-09 ***
## OLDCLAIM -7.966e-06 4.927e-06 -1.617 0.105879
## CLM_FREQ 1.887e-01 3.675e-02 5.134 2.84e-07 ***
## REVOKED 8.491e-01 1.173e-01 7.237 4.57e-13 ***
## MVR_PTS 1.023e-01 1.804e-02 5.674 1.39e-08 ***
## URBANICITY 2.392e+00 1.444e-01 16.568 < 2e-16 ***
## Clerical 1.788e-01 1.175e-01 1.523 0.127830
## Doctor -5.196e-01 2.979e-01 -1.744 0.081188 .
## Manager -9.267e-01 1.462e-01 -6.339 2.31e-10 ***
## Professional -2.251e-01 1.203e-01 -1.871 0.061292 .
## `Panel Truck` 6.411e-01 1.856e-01 3.453 0.000554 ***
## Pickup 4.211e-01 1.258e-01 3.348 0.000815 ***
## `Sports Car` 8.996e-01 1.408e-01 6.389 1.66e-10 ***
## Van 5.873e-01 1.536e-01 3.824 0.000132 ***
## z_SUV 6.898e-01 1.090e-01 6.329 2.46e-10 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4392.9 on 4870 degrees of freedom
## AIC: 4446.9
##
## Number of Fisher Scoring iterations: 5Home Value has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train2$TARGET_FLAG ~ KIDSDRIV + HOMEKIDS + YOJ +
## INCOME + PARENT1 + MSTATUS + EDUCATION + TRAVTIME + CAR_USE +
## BLUEBOOK + TIF + OLDCLAIM + CLM_FREQ + REVOKED + MVR_PTS +
## URBANICITY + Clerical + Doctor + Manager + Professional +
## `Panel Truck` + Pickup + `Sports Car` + Van + z_SUV, family = binomial(link = "logit"),
## data = train2)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2451 -0.7196 -0.4025 0.6311 3.1149
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.231e+00 2.360e-01 -13.694 < 2e-16 ***
## KIDSDRIV 3.218e-01 7.813e-02 4.119 3.81e-05 ***
## HOMEKIDS 6.865e-02 4.314e-02 1.591 0.111537
## YOJ -1.889e-02 1.005e-02 -1.880 0.060135 .
## INCOME -5.273e-06 1.207e-06 -4.370 1.25e-05 ***
## PARENT1 4.033e-01 1.413e-01 2.854 0.004323 **
## MSTATUS -6.144e-01 9.421e-02 -6.522 6.96e-11 ***
## EDUCATION -1.170e-01 4.306e-02 -2.718 0.006570 **
## TRAVTIME 1.595e-02 2.426e-03 6.577 4.81e-11 ***
## CAR_USE 8.097e-01 9.663e-02 8.380 < 2e-16 ***
## BLUEBOOK -2.593e-05 6.174e-06 -4.199 2.68e-05 ***
## TIF -5.470e-02 9.398e-03 -5.821 5.85e-09 ***
## OLDCLAIM -7.926e-06 4.922e-06 -1.610 0.107338
## CLM_FREQ 1.868e-01 3.670e-02 5.088 3.61e-07 ***
## REVOKED 8.471e-01 1.173e-01 7.224 5.04e-13 ***
## MVR_PTS 1.028e-01 1.802e-02 5.706 1.16e-08 ***
## URBANICITY 2.389e+00 1.442e-01 16.565 < 2e-16 ***
## Clerical 1.684e-01 1.170e-01 1.439 0.150220
## Doctor -5.146e-01 2.982e-01 -1.726 0.084364 .
## Manager -9.267e-01 1.463e-01 -6.334 2.38e-10 ***
## Professional -2.222e-01 1.203e-01 -1.847 0.064684 .
## `Panel Truck` 6.466e-01 1.856e-01 3.484 0.000495 ***
## Pickup 4.181e-01 1.257e-01 3.325 0.000885 ***
## `Sports Car` 8.997e-01 1.407e-01 6.393 1.63e-10 ***
## Van 5.901e-01 1.536e-01 3.841 0.000123 ***
## z_SUV 6.900e-01 1.090e-01 6.333 2.40e-10 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4394.4 on 4871 degrees of freedom
## AIC: 4446.4
##
## Number of Fisher Scoring iterations: 5Clerical has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train2$TARGET_FLAG ~ KIDSDRIV + HOMEKIDS + YOJ +
## INCOME + PARENT1 + MSTATUS + EDUCATION + TRAVTIME + CAR_USE +
## BLUEBOOK + TIF + OLDCLAIM + CLM_FREQ + REVOKED + MVR_PTS +
## URBANICITY + Doctor + Manager + Professional + `Panel Truck` +
## Pickup + `Sports Car` + Van + z_SUV, family = binomial(link = "logit"),
## data = train2)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2635 -0.7197 -0.4021 0.6310 3.1090
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.172e+00 2.319e-01 -13.675 < 2e-16 ***
## KIDSDRIV 3.197e-01 7.812e-02 4.093 4.27e-05 ***
## HOMEKIDS 6.801e-02 4.313e-02 1.577 0.114787
## YOJ -1.611e-02 9.850e-03 -1.635 0.102014
## INCOME -5.324e-06 1.205e-06 -4.417 9.99e-06 ***
## PARENT1 4.110e-01 1.412e-01 2.911 0.003605 **
## MSTATUS -6.136e-01 9.415e-02 -6.518 7.14e-11 ***
## EDUCATION -1.334e-01 4.146e-02 -3.218 0.001290 **
## TRAVTIME 1.585e-02 2.423e-03 6.540 6.14e-11 ***
## CAR_USE 7.671e-01 9.176e-02 8.361 < 2e-16 ***
## BLUEBOOK -2.594e-05 6.174e-06 -4.201 2.66e-05 ***
## TIF -5.433e-02 9.390e-03 -5.786 7.22e-09 ***
## OLDCLAIM -7.944e-06 4.923e-06 -1.614 0.106626
## CLM_FREQ 1.869e-01 3.669e-02 5.094 3.51e-07 ***
## REVOKED 8.494e-01 1.172e-01 7.246 4.28e-13 ***
## MVR_PTS 1.033e-01 1.801e-02 5.738 9.56e-09 ***
## URBANICITY 2.375e+00 1.439e-01 16.502 < 2e-16 ***
## Doctor -5.227e-01 2.980e-01 -1.754 0.079421 .
## Manager -9.567e-01 1.447e-01 -6.612 3.80e-11 ***
## Professional -2.608e-01 1.171e-01 -2.228 0.025909 *
## `Panel Truck` 6.813e-01 1.840e-01 3.702 0.000214 ***
## Pickup 4.369e-01 1.249e-01 3.497 0.000471 ***
## `Sports Car` 8.955e-01 1.406e-01 6.367 1.92e-10 ***
## Van 6.085e-01 1.530e-01 3.978 6.95e-05 ***
## z_SUV 6.881e-01 1.089e-01 6.318 2.65e-10 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4396.4 on 4872 degrees of freedom
## AIC: 4446.4
##
## Number of Fisher Scoring iterations: 5HomeKIDS has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train2$TARGET_FLAG ~ KIDSDRIV + YOJ + INCOME +
## PARENT1 + MSTATUS + EDUCATION + TRAVTIME + CAR_USE + BLUEBOOK +
## TIF + OLDCLAIM + CLM_FREQ + REVOKED + MVR_PTS + URBANICITY +
## Doctor + Manager + Professional + `Panel Truck` + Pickup +
## `Sports Car` + Van + z_SUV, family = binomial(link = "logit"),
## data = train2)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2785 -0.7210 -0.4025 0.6308 3.0979
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.164e+00 2.317e-01 -13.657 < 2e-16 ***
## KIDSDRIV 3.691e-01 7.167e-02 5.150 2.60e-07 ***
## YOJ -1.457e-02 9.802e-03 -1.487 0.137034
## INCOME -5.392e-06 1.205e-06 -4.474 7.68e-06 ***
## PARENT1 5.210e-01 1.229e-01 4.241 2.23e-05 ***
## MSTATUS -5.704e-01 8.986e-02 -6.348 2.19e-10 ***
## EDUCATION -1.391e-01 4.128e-02 -3.370 0.000753 ***
## TRAVTIME 1.572e-02 2.421e-03 6.495 8.30e-11 ***
## CAR_USE 7.669e-01 9.173e-02 8.360 < 2e-16 ***
## BLUEBOOK -2.631e-05 6.171e-06 -4.264 2.01e-05 ***
## TIF -5.393e-02 9.382e-03 -5.748 9.03e-09 ***
## OLDCLAIM -8.045e-06 4.923e-06 -1.634 0.102182
## CLM_FREQ 1.862e-01 3.668e-02 5.076 3.85e-07 ***
## REVOKED 8.556e-01 1.171e-01 7.308 2.71e-13 ***
## MVR_PTS 1.041e-01 1.800e-02 5.785 7.23e-09 ***
## URBANICITY 2.375e+00 1.439e-01 16.503 < 2e-16 ***
## Doctor -5.278e-01 2.980e-01 -1.771 0.076506 .
## Manager -9.682e-01 1.445e-01 -6.698 2.11e-11 ***
## Professional -2.701e-01 1.169e-01 -2.310 0.020898 *
## `Panel Truck` 6.870e-01 1.839e-01 3.735 0.000188 ***
## Pickup 4.369e-01 1.249e-01 3.498 0.000469 ***
## `Sports Car` 8.999e-01 1.405e-01 6.403 1.53e-10 ***
## Van 6.137e-01 1.529e-01 4.014 5.98e-05 ***
## z_SUV 6.925e-01 1.088e-01 6.363 1.97e-10 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4398.9 on 4873 degrees of freedom
## AIC: 4446.9
##
## Number of Fisher Scoring iterations: 5Year On the Job has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train2$TARGET_FLAG ~ KIDSDRIV + INCOME + PARENT1 +
## MSTATUS + EDUCATION + TRAVTIME + CAR_USE + BLUEBOOK + TIF +
## OLDCLAIM + CLM_FREQ + REVOKED + MVR_PTS + URBANICITY + Doctor +
## Manager + Professional + `Panel Truck` + Pickup + `Sports Car` +
## Van + z_SUV, family = binomial(link = "logit"), data = train2)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2850 -0.7212 -0.4011 0.6410 3.0893
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.272e+00 2.205e-01 -14.835 < 2e-16 ***
## KIDSDRIV 3.686e-01 7.164e-02 5.146 2.66e-07 ***
## INCOME -5.917e-06 1.158e-06 -5.108 3.26e-07 ***
## PARENT1 5.127e-01 1.226e-01 4.180 2.92e-05 ***
## MSTATUS -5.911e-01 8.879e-02 -6.658 2.77e-11 ***
## EDUCATION -1.315e-01 4.095e-02 -3.212 0.001319 **
## TRAVTIME 1.563e-02 2.417e-03 6.466 1.01e-10 ***
## CAR_USE 7.683e-01 9.169e-02 8.380 < 2e-16 ***
## BLUEBOOK -2.661e-05 6.172e-06 -4.312 1.62e-05 ***
## TIF -5.413e-02 9.377e-03 -5.773 7.81e-09 ***
## OLDCLAIM -8.308e-06 4.918e-06 -1.690 0.091123 .
## CLM_FREQ 1.869e-01 3.666e-02 5.098 3.44e-07 ***
## REVOKED 8.549e-01 1.170e-01 7.305 2.77e-13 ***
## MVR_PTS 1.047e-01 1.798e-02 5.822 5.83e-09 ***
## URBANICITY 2.365e+00 1.436e-01 16.476 < 2e-16 ***
## Doctor -5.291e-01 2.981e-01 -1.775 0.075885 .
## Manager -9.769e-01 1.445e-01 -6.762 1.36e-11 ***
## Professional -2.791e-01 1.168e-01 -2.390 0.016863 *
## `Panel Truck` 6.991e-01 1.838e-01 3.803 0.000143 ***
## Pickup 4.374e-01 1.248e-01 3.504 0.000458 ***
## `Sports Car` 9.153e-01 1.401e-01 6.533 6.44e-11 ***
## Van 6.198e-01 1.529e-01 4.055 5.02e-05 ***
## z_SUV 6.976e-01 1.087e-01 6.417 1.39e-10 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4401.1 on 4874 degrees of freedom
## AIC: 4447.1
##
## Number of Fisher Scoring iterations: 5OldClaim has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train2$TARGET_FLAG ~ KIDSDRIV + INCOME + PARENT1 +
## MSTATUS + EDUCATION + TRAVTIME + CAR_USE + BLUEBOOK + TIF +
## CLM_FREQ + REVOKED + MVR_PTS + URBANICITY + Doctor + Manager +
## Professional + `Panel Truck` + Pickup + `Sports Car` + Van +
## z_SUV, family = binomial(link = "logit"), data = train2)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2625 -0.7246 -0.4026 0.6340 3.0865
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.271e+00 2.204e-01 -14.843 < 2e-16 ***
## KIDSDRIV 3.725e-01 7.151e-02 5.208 1.91e-07 ***
## INCOME -5.908e-06 1.158e-06 -5.103 3.35e-07 ***
## PARENT1 5.141e-01 1.226e-01 4.195 2.73e-05 ***
## MSTATUS -5.892e-01 8.875e-02 -6.639 3.16e-11 ***
## EDUCATION -1.309e-01 4.093e-02 -3.199 0.001381 **
## TRAVTIME 1.574e-02 2.415e-03 6.519 7.07e-11 ***
## CAR_USE 7.675e-01 9.169e-02 8.371 < 2e-16 ***
## BLUEBOOK -2.672e-05 6.167e-06 -4.333 1.47e-05 ***
## TIF -5.440e-02 9.370e-03 -5.805 6.42e-09 ***
## CLM_FREQ 1.592e-01 3.288e-02 4.841 1.29e-06 ***
## REVOKED 7.627e-01 1.039e-01 7.343 2.09e-13 ***
## MVR_PTS 1.017e-01 1.788e-02 5.688 1.28e-08 ***
## URBANICITY 2.364e+00 1.435e-01 16.479 < 2e-16 ***
## Doctor -5.236e-01 2.976e-01 -1.760 0.078451 .
## Manager -9.757e-01 1.444e-01 -6.756 1.42e-11 ***
## Professional -2.766e-01 1.167e-01 -2.370 0.017803 *
## `Panel Truck` 7.017e-01 1.837e-01 3.821 0.000133 ***
## Pickup 4.384e-01 1.249e-01 3.511 0.000446 ***
## `Sports Car` 9.136e-01 1.401e-01 6.522 6.93e-11 ***
## Van 6.211e-01 1.528e-01 4.065 4.81e-05 ***
## z_SUV 7.009e-01 1.086e-01 6.452 1.10e-10 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4404.0 on 4875 degrees of freedom
## AIC: 4448
##
## Number of Fisher Scoring iterations: 5Doctor has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train2$TARGET_FLAG ~ KIDSDRIV + INCOME + PARENT1 +
## MSTATUS + EDUCATION + TRAVTIME + CAR_USE + BLUEBOOK + TIF +
## CLM_FREQ + REVOKED + MVR_PTS + URBANICITY + Manager + Professional +
## `Panel Truck` + Pickup + `Sports Car` + Van + z_SUV, family = binomial(link = "logit"),
## data = train2)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2685 -0.7237 -0.4043 0.6322 3.0893
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.261e+00 2.203e-01 -14.800 < 2e-16 ***
## KIDSDRIV 3.747e-01 7.157e-02 5.236 1.64e-07 ***
## INCOME -6.114e-06 1.152e-06 -5.309 1.10e-07 ***
## PARENT1 5.155e-01 1.225e-01 4.208 2.58e-05 ***
## MSTATUS -5.866e-01 8.870e-02 -6.613 3.76e-11 ***
## EDUCATION -1.485e-01 3.979e-02 -3.732 0.000190 ***
## TRAVTIME 1.576e-02 2.415e-03 6.524 6.85e-11 ***
## CAR_USE 7.903e-01 9.092e-02 8.693 < 2e-16 ***
## BLUEBOOK -2.679e-05 6.160e-06 -4.349 1.37e-05 ***
## TIF -5.410e-02 9.372e-03 -5.772 7.81e-09 ***
## CLM_FREQ 1.589e-01 3.288e-02 4.834 1.34e-06 ***
## REVOKED 7.661e-01 1.038e-01 7.381 1.57e-13 ***
## MVR_PTS 1.018e-01 1.789e-02 5.691 1.26e-08 ***
## URBANICITY 2.366e+00 1.436e-01 16.473 < 2e-16 ***
## Manager -9.406e-01 1.433e-01 -6.566 5.16e-11 ***
## Professional -2.498e-01 1.159e-01 -2.155 0.031186 *
## `Panel Truck` 7.148e-01 1.836e-01 3.893 9.89e-05 ***
## Pickup 4.341e-01 1.248e-01 3.477 0.000506 ***
## `Sports Car` 9.109e-01 1.399e-01 6.510 7.50e-11 ***
## Van 6.195e-01 1.526e-01 4.059 4.92e-05 ***
## z_SUV 6.987e-01 1.086e-01 6.435 1.24e-10 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 4407.4 on 4876 degrees of freedom
## AIC: 4449.4
##
## Number of Fisher Scoring iterations: 5## (Intercept) KIDSDRIV INCOME PARENT1 MSTATUS
## -4.778999e-01 5.492667e-02 -8.961659e-07 7.555801e-02 -8.598043e-02
## EDUCATION TRAVTIME CAR_USE BLUEBOOK TIF
## -2.176489e-02 2.309464e-03 1.158372e-01 -3.926466e-06 -7.929253e-03
## CLM_FREQ REVOKED MVR_PTS URBANICITY Manager
## 2.329275e-02 1.122878e-01 1.492372e-02 3.467393e-01 -1.378694e-01
## Professional `Panel Truck` Pickup `Sports Car` Van
## -3.661444e-02 1.047719e-01 6.362690e-02 1.335134e-01 9.080778e-02
## z_SUV
## 1.024094e-01The variables that are likely to result in a customer having a claim are: having more kids who are driving, being a single parent, traveling a longer distance to work, using a commercial vehicle, the more claims a customer had in the last 5 years, a customer having his/her license revoked, a customer having points, being in an urban area, driving a panel truck, pickup, sports car van or SUV. The following variables are likely to result in a person less likely to have a claim: higher income, being married, having a higher education, a car of more value, being on the job for more years, being a manager, and being a professional.
Prediction from Model 2
## [1] 0.33## [1] 0.7974893The cutoff associated with a point the farthest distance from the ROC curve is 0.33. I will use 0.33 as the cutoff for making predictions. A value above 0.33 will be assigned a target of one and a value below 0.33 wil be assigned a value of zero. The area under the curve is 0.80.
Confusion Matrix
## true
## pred 0 1
## 0 1875 268
## 1 531 590The model predicted 1,875 0’s that were actually 0. The model predicted 268 0’s that were actually 1.
The model predicted 531 1’s that were actaully 0. The model predicted 590 1’s that were actually 1.
## accuracy error precision sensitivity specificity f1
## 1 0.7552083 0.2447917 0.4121785 0.7793017 0.6876457 0.5391804Model3
Run with a different seed for separating the data
Creating a test and training set
Backward Regression
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ KIDSDRIV + YOJ + MSTATUS +
## TRAVTIME + CAR_USE + BLUEBOOK + TIF + REVOKED + MVR_PTS +
## CAR_AGE + URBANICITY + Clerical + Doctor + `Home Maker` +
## Lawyer + Manager + Professional + `z_Blue Collar` + `Panel Truck` +
## Pickup + `Sports Car` + Van + homekids_age + income_homeval_educ +
## sex_redcar_suv + oldclaim_freq, family = binomial(link = "logit"),
## data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.0398 -0.8044 -0.7526 1.4979 1.8606
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -8.638e-01 2.030e-01 -4.254 2.1e-05 ***
## KIDSDRIV 6.414e-02 7.477e-02 0.858 0.390998
## YOJ 9.212e-03 9.523e-03 0.967 0.333395
## MSTATUS 1.179e-02 6.766e-02 0.174 0.861699
## TRAVTIME 2.469e-03 2.073e-03 1.191 0.233508
## CAR_USE -9.222e-02 1.016e-01 -0.907 0.364196
## BLUEBOOK -1.638e-06 4.921e-06 -0.333 0.739158
## TIF -3.437e-03 7.936e-03 -0.433 0.664965
## REVOKED 6.899e-02 1.083e-01 0.637 0.524248
## MVR_PTS -1.283e-02 1.604e-02 -0.800 0.423719
## CAR_AGE -4.876e-03 7.434e-03 -0.656 0.511868
## URBANICITY -1.015e-01 8.743e-02 -1.161 0.245775
## Clerical -4.687e-01 1.342e-01 -3.493 0.000477 ***
## Doctor -3.298e-01 2.279e-01 -1.447 0.147837
## `Home Maker` -4.241e-01 1.595e-01 -2.659 0.007833 **
## Lawyer -2.794e-01 1.538e-01 -1.816 0.069363 .
## Manager -1.584e-01 1.357e-01 -1.167 0.243126
## Professional -2.864e-01 1.290e-01 -2.221 0.026363 *
## `z_Blue Collar` -1.617e-01 1.194e-01 -1.355 0.175520
## `Panel Truck` -2.424e-03 1.653e-01 -0.015 0.988304
## Pickup -9.405e-02 1.061e-01 -0.886 0.375444
## `Sports Car` 1.231e-01 1.095e-01 1.124 0.260881
## Van -8.620e-02 1.341e-01 -0.643 0.520225
## homekids_age -1.729e-04 9.151e-04 -0.189 0.850109
## income_homeval_educ 1.061e-07 8.569e-08 1.238 0.215830
## sex_redcar_suv 4.173e-02 5.740e-02 0.727 0.467188
## oldclaim_freq -6.457e-07 4.297e-06 -0.150 0.880562
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5627.8 on 4870 degrees of freedom
## AIC: 5681.8
##
## Number of Fisher Scoring iterations: 4Panel Truck has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ KIDSDRIV + YOJ + MSTATUS +
## TRAVTIME + CAR_USE + BLUEBOOK + TIF + REVOKED + MVR_PTS +
## CAR_AGE + URBANICITY + Clerical + Doctor + `Home Maker` +
## Lawyer + Manager + Professional + `z_Blue Collar` + Pickup +
## `Sports Car` + Van + homekids_age + income_homeval_educ +
## sex_redcar_suv + oldclaim_freq, family = binomial(link = "logit"),
## data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.0397 -0.8044 -0.7527 1.4982 1.8610
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -8.638e-01 2.030e-01 -4.254 2.1e-05 ***
## KIDSDRIV 6.412e-02 7.476e-02 0.858 0.391064
## YOJ 9.208e-03 9.519e-03 0.967 0.333393
## MSTATUS 1.181e-02 6.765e-02 0.175 0.861472
## TRAVTIME 2.469e-03 2.072e-03 1.191 0.233537
## CAR_USE -9.281e-02 9.316e-02 -0.996 0.319143
## BLUEBOOK -1.666e-06 4.534e-06 -0.368 0.713241
## TIF -3.437e-03 7.936e-03 -0.433 0.664917
## REVOKED 6.901e-02 1.083e-01 0.637 0.524136
## MVR_PTS -1.283e-02 1.604e-02 -0.800 0.423781
## CAR_AGE -4.877e-03 7.434e-03 -0.656 0.511844
## URBANICITY -1.015e-01 8.739e-02 -1.162 0.245425
## Clerical -4.687e-01 1.342e-01 -3.494 0.000476 ***
## Doctor -3.295e-01 2.267e-01 -1.453 0.146169
## `Home Maker` -4.241e-01 1.595e-01 -2.659 0.007834 **
## Lawyer -2.792e-01 1.529e-01 -1.825 0.067971 .
## Manager -1.583e-01 1.355e-01 -1.169 0.242586
## Professional -2.863e-01 1.287e-01 -2.225 0.026115 *
## `z_Blue Collar` -1.613e-01 1.162e-01 -1.389 0.164984
## Pickup -9.353e-02 1.001e-01 -0.935 0.350028
## `Sports Car` 1.233e-01 1.087e-01 1.134 0.256604
## Van -8.543e-02 1.232e-01 -0.693 0.488075
## homekids_age -1.725e-04 9.146e-04 -0.189 0.850407
## income_homeval_educ 1.060e-07 8.558e-08 1.239 0.215504
## sex_redcar_suv 4.193e-02 5.575e-02 0.752 0.452006
## oldclaim_freq -6.458e-07 4.297e-06 -0.150 0.880546
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5627.8 on 4871 degrees of freedom
## AIC: 5679.8
##
## Number of Fisher Scoring iterations: 4oldclaim_freq has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ KIDSDRIV + YOJ + MSTATUS +
## TRAVTIME + CAR_USE + BLUEBOOK + TIF + REVOKED + MVR_PTS +
## CAR_AGE + URBANICITY + Clerical + Doctor + `Home Maker` +
## Lawyer + Manager + Professional + `z_Blue Collar` + Pickup +
## `Sports Car` + Van + homekids_age + income_homeval_educ +
## sex_redcar_suv, family = binomial(link = "logit"), data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.0366 -0.8041 -0.7526 1.4989 1.8615
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -8.637e-01 2.030e-01 -4.254 2.1e-05 ***
## KIDSDRIV 6.417e-02 7.476e-02 0.858 0.390727
## YOJ 9.184e-03 9.518e-03 0.965 0.334559
## MSTATUS 1.208e-02 6.763e-02 0.179 0.858237
## TRAVTIME 2.464e-03 2.072e-03 1.189 0.234468
## CAR_USE -9.295e-02 9.316e-02 -0.998 0.318395
## BLUEBOOK -1.656e-06 4.534e-06 -0.365 0.714981
## TIF -3.426e-03 7.936e-03 -0.432 0.665960
## REVOKED 6.226e-02 9.866e-02 0.631 0.527994
## MVR_PTS -1.342e-02 1.556e-02 -0.862 0.388550
## CAR_AGE -4.883e-03 7.434e-03 -0.657 0.511308
## URBANICITY -1.029e-01 8.690e-02 -1.184 0.236272
## Clerical -4.688e-01 1.342e-01 -3.494 0.000475 ***
## Doctor -3.291e-01 2.267e-01 -1.452 0.146636
## `Home Maker` -4.241e-01 1.595e-01 -2.659 0.007831 **
## Lawyer -2.784e-01 1.529e-01 -1.821 0.068566 .
## Manager -1.580e-01 1.355e-01 -1.166 0.243455
## Professional -2.857e-01 1.286e-01 -2.221 0.026353 *
## `z_Blue Collar` -1.611e-01 1.162e-01 -1.386 0.165597
## Pickup -9.317e-02 1.001e-01 -0.931 0.351737
## `Sports Car` 1.228e-01 1.087e-01 1.130 0.258386
## Van -8.564e-02 1.232e-01 -0.695 0.486980
## homekids_age -1.696e-04 9.144e-04 -0.185 0.852854
## income_homeval_educ 1.062e-07 8.556e-08 1.241 0.214462
## sex_redcar_suv 4.192e-02 5.575e-02 0.752 0.452118
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5627.8 on 4872 degrees of freedom
## AIC: 5677.8
##
## Number of Fisher Scoring iterations: 4MSTATUS has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ KIDSDRIV + YOJ + TRAVTIME +
## CAR_USE + BLUEBOOK + TIF + REVOKED + MVR_PTS + CAR_AGE +
## URBANICITY + Clerical + Doctor + `Home Maker` + Lawyer +
## Manager + Professional + `z_Blue Collar` + Pickup + `Sports Car` +
## Van + homekids_age + income_homeval_educ + sex_redcar_suv,
## family = binomial(link = "logit"), data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.0399 -0.8039 -0.7530 1.5001 1.8630
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -8.584e-01 2.008e-01 -4.274 1.92e-05 ***
## KIDSDRIV 6.457e-02 7.473e-02 0.864 0.387540
## YOJ 9.459e-03 9.395e-03 1.007 0.314017
## TRAVTIME 2.466e-03 2.072e-03 1.190 0.234094
## CAR_USE -9.334e-02 9.313e-02 -1.002 0.316231
## BLUEBOOK -1.678e-06 4.532e-06 -0.370 0.711243
## TIF -3.416e-03 7.936e-03 -0.430 0.666853
## REVOKED 6.138e-02 9.854e-02 0.623 0.533344
## MVR_PTS -1.354e-02 1.554e-02 -0.871 0.383680
## CAR_AGE -4.946e-03 7.426e-03 -0.666 0.505391
## URBANICITY -1.027e-01 8.689e-02 -1.182 0.237226
## Clerical -4.693e-01 1.341e-01 -3.499 0.000467 ***
## Doctor -3.315e-01 2.263e-01 -1.465 0.142871
## `Home Maker` -4.238e-01 1.595e-01 -2.657 0.007875 **
## Lawyer -2.797e-01 1.527e-01 -1.832 0.066925 .
## Manager -1.589e-01 1.354e-01 -1.173 0.240641
## Professional -2.864e-01 1.286e-01 -2.227 0.025933 *
## `z_Blue Collar` -1.615e-01 1.161e-01 -1.391 0.164281
## Pickup -9.327e-02 1.001e-01 -0.932 0.351212
## `Sports Car` 1.230e-01 1.087e-01 1.132 0.257587
## Van -8.546e-02 1.232e-01 -0.694 0.487872
## homekids_age -1.683e-04 9.144e-04 -0.184 0.853980
## income_homeval_educ 1.070e-07 8.545e-08 1.252 0.210408
## sex_redcar_suv 4.182e-02 5.575e-02 0.750 0.453235
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5627.9 on 4873 degrees of freedom
## AIC: 5675.9
##
## Number of Fisher Scoring iterations: 4Homekidsage has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ KIDSDRIV + YOJ + TRAVTIME +
## CAR_USE + BLUEBOOK + TIF + REVOKED + MVR_PTS + CAR_AGE +
## URBANICITY + Clerical + Doctor + `Home Maker` + Lawyer +
## Manager + Professional + `z_Blue Collar` + Pickup + `Sports Car` +
## Van + income_homeval_educ + sex_redcar_suv, family = binomial(link = "logit"),
## data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.0402 -0.8041 -0.7528 1.4982 1.8643
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -8.605e-01 2.005e-01 -4.292 1.77e-05 ***
## KIDSDRIV 5.716e-02 6.294e-02 0.908 0.363843
## YOJ 9.160e-03 9.251e-03 0.990 0.322093
## TRAVTIME 2.475e-03 2.072e-03 1.195 0.232238
## CAR_USE -9.305e-02 9.312e-02 -0.999 0.317677
## BLUEBOOK -1.654e-06 4.530e-06 -0.365 0.714985
## TIF -3.426e-03 7.936e-03 -0.432 0.665915
## REVOKED 6.088e-02 9.850e-02 0.618 0.536536
## MVR_PTS -1.359e-02 1.554e-02 -0.874 0.381967
## CAR_AGE -4.897e-03 7.421e-03 -0.660 0.509320
## URBANICITY -1.026e-01 8.689e-02 -1.181 0.237576
## Clerical -4.683e-01 1.340e-01 -3.495 0.000475 ***
## Doctor -3.290e-01 2.259e-01 -1.457 0.145214
## `Home Maker` -4.230e-01 1.594e-01 -2.653 0.007970 **
## Lawyer -2.770e-01 1.520e-01 -1.823 0.068325 .
## Manager -1.567e-01 1.349e-01 -1.162 0.245354
## Professional -2.841e-01 1.280e-01 -2.220 0.026437 *
## `z_Blue Collar` -1.597e-01 1.157e-01 -1.380 0.167499
## Pickup -9.349e-02 1.000e-01 -0.935 0.350020
## `Sports Car` 1.221e-01 1.085e-01 1.125 0.260651
## Van -8.542e-02 1.232e-01 -0.693 0.488030
## income_homeval_educ 1.079e-07 8.531e-08 1.265 0.205965
## sex_redcar_suv 4.093e-02 5.554e-02 0.737 0.461190
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5627.9 on 4874 degrees of freedom
## AIC: 5673.9
##
## Number of Fisher Scoring iterations: 4BLUEBOOK has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ KIDSDRIV + YOJ + TRAVTIME +
## CAR_USE + TIF + REVOKED + MVR_PTS + CAR_AGE + URBANICITY +
## Clerical + Doctor + `Home Maker` + Lawyer + Manager + Professional +
## `z_Blue Collar` + Pickup + `Sports Car` + Van + income_homeval_educ +
## sex_redcar_suv, family = binomial(link = "logit"), data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.0455 -0.8039 -0.7527 1.5009 1.8699
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -8.816e-01 1.920e-01 -4.592 4.39e-06 ***
## KIDSDRIV 5.690e-02 6.294e-02 0.904 0.36594
## YOJ 8.939e-03 9.232e-03 0.968 0.33292
## TRAVTIME 2.477e-03 2.072e-03 1.196 0.23187
## CAR_USE -1.006e-01 9.074e-02 -1.109 0.26743
## TIF -3.414e-03 7.936e-03 -0.430 0.66703
## REVOKED 6.134e-02 9.849e-02 0.623 0.53339
## MVR_PTS -1.339e-02 1.553e-02 -0.862 0.38852
## CAR_AGE -4.889e-03 7.422e-03 -0.659 0.51003
## URBANICITY -1.027e-01 8.689e-02 -1.182 0.23709
## Clerical -4.706e-01 1.339e-01 -3.515 0.00044 ***
## Doctor -3.284e-01 2.258e-01 -1.454 0.14597
## `Home Maker` -4.251e-01 1.593e-01 -2.668 0.00762 **
## Lawyer -2.779e-01 1.519e-01 -1.829 0.06737 .
## Manager -1.602e-01 1.345e-01 -1.191 0.23359
## Professional -2.886e-01 1.274e-01 -2.265 0.02349 *
## `z_Blue Collar` -1.596e-01 1.157e-01 -1.379 0.16784
## Pickup -8.491e-02 9.719e-02 -0.874 0.38235
## `Sports Car` 1.281e-01 1.073e-01 1.194 0.23233
## Van -8.668e-02 1.231e-01 -0.704 0.48137
## income_homeval_educ 1.004e-07 8.279e-08 1.212 0.22544
## sex_redcar_suv 4.424e-02 5.483e-02 0.807 0.41969
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5628.0 on 4875 degrees of freedom
## AIC: 5672
##
## Number of Fisher Scoring iterations: 4TIF has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ KIDSDRIV + YOJ + TRAVTIME +
## CAR_USE + REVOKED + MVR_PTS + CAR_AGE + URBANICITY + Clerical +
## Doctor + `Home Maker` + Lawyer + Manager + Professional +
## `z_Blue Collar` + Pickup + `Sports Car` + Van + income_homeval_educ +
## sex_redcar_suv, family = binomial(link = "logit"), data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.0442 -0.8039 -0.7533 1.5004 1.8658
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -9.008e-01 1.868e-01 -4.822 1.42e-06 ***
## KIDSDRIV 5.744e-02 6.293e-02 0.913 0.36137
## YOJ 8.923e-03 9.231e-03 0.967 0.33370
## TRAVTIME 2.477e-03 2.071e-03 1.196 0.23172
## CAR_USE -1.006e-01 9.074e-02 -1.109 0.26761
## REVOKED 6.258e-02 9.845e-02 0.636 0.52501
## MVR_PTS -1.309e-02 1.551e-02 -0.844 0.39891
## CAR_AGE -4.923e-03 7.421e-03 -0.663 0.50704
## URBANICITY -1.031e-01 8.688e-02 -1.187 0.23522
## Clerical -4.706e-01 1.339e-01 -3.515 0.00044 ***
## Doctor -3.262e-01 2.258e-01 -1.445 0.14858
## `Home Maker` -4.229e-01 1.592e-01 -2.656 0.00790 **
## Lawyer -2.767e-01 1.519e-01 -1.822 0.06853 .
## Manager -1.604e-01 1.345e-01 -1.193 0.23304
## Professional -2.878e-01 1.274e-01 -2.259 0.02389 *
## `z_Blue Collar` -1.589e-01 1.157e-01 -1.373 0.16960
## Pickup -8.465e-02 9.718e-02 -0.871 0.38374
## `Sports Car` 1.281e-01 1.073e-01 1.194 0.23230
## Van -8.716e-02 1.231e-01 -0.708 0.47892
## income_homeval_educ 1.009e-07 8.279e-08 1.218 0.22315
## sex_redcar_suv 4.428e-02 5.482e-02 0.808 0.41923
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5628.2 on 4876 degrees of freedom
## AIC: 5670.2
##
## Number of Fisher Scoring iterations: 4REVOKED has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ KIDSDRIV + YOJ + TRAVTIME +
## CAR_USE + MVR_PTS + CAR_AGE + URBANICITY + Clerical + Doctor +
## `Home Maker` + Lawyer + Manager + Professional + `z_Blue Collar` +
## Pickup + `Sports Car` + Van + income_homeval_educ + sex_redcar_suv,
## family = binomial(link = "logit"), data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.0302 -0.8039 -0.7536 1.5016 1.8617
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -8.968e-01 1.867e-01 -4.804 1.55e-06 ***
## KIDSDRIV 5.884e-02 6.288e-02 0.936 0.349464
## YOJ 8.880e-03 9.229e-03 0.962 0.335981
## TRAVTIME 2.470e-03 2.071e-03 1.193 0.233001
## CAR_USE -1.000e-01 9.071e-02 -1.102 0.270260
## MVR_PTS -1.284e-02 1.551e-02 -0.828 0.407894
## CAR_AGE -4.888e-03 7.420e-03 -0.659 0.510034
## URBANICITY -9.744e-02 8.640e-02 -1.128 0.259423
## Clerical -4.719e-01 1.338e-01 -3.526 0.000422 ***
## Doctor -3.285e-01 2.257e-01 -1.455 0.145553
## `Home Maker` -4.256e-01 1.591e-01 -2.674 0.007491 **
## Lawyer -2.775e-01 1.519e-01 -1.827 0.067665 .
## Manager -1.634e-01 1.344e-01 -1.216 0.224106
## Professional -2.891e-01 1.274e-01 -2.270 0.023216 *
## `z_Blue Collar` -1.602e-01 1.157e-01 -1.385 0.166158
## Pickup -8.396e-02 9.717e-02 -0.864 0.387535
## `Sports Car` 1.282e-01 1.073e-01 1.195 0.232108
## Van -8.678e-02 1.231e-01 -0.705 0.480791
## income_homeval_educ 9.851e-08 8.270e-08 1.191 0.233618
## sex_redcar_suv 4.572e-02 5.477e-02 0.835 0.403864
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5628.6 on 4877 degrees of freedom
## AIC: 5668.6
##
## Number of Fisher Scoring iterations: 4CAR_AGE has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ KIDSDRIV + YOJ + TRAVTIME +
## CAR_USE + MVR_PTS + URBANICITY + Clerical + Doctor + `Home Maker` +
## Lawyer + Manager + Professional + `z_Blue Collar` + Pickup +
## `Sports Car` + Van + income_homeval_educ + sex_redcar_suv,
## family = binomial(link = "logit"), data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.0333 -0.8043 -0.7537 1.5048 1.8667
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -9.231e-01 1.824e-01 -5.060 4.2e-07 ***
## KIDSDRIV 5.844e-02 6.288e-02 0.929 0.352690
## YOJ 8.844e-03 9.230e-03 0.958 0.338016
## TRAVTIME 2.461e-03 2.071e-03 1.188 0.234653
## CAR_USE -1.020e-01 9.069e-02 -1.125 0.260702
## MVR_PTS -1.299e-02 1.551e-02 -0.837 0.402427
## URBANICITY -9.952e-02 8.635e-02 -1.153 0.249107
## Clerical -4.643e-01 1.334e-01 -3.481 0.000499 ***
## Doctor -3.315e-01 2.256e-01 -1.469 0.141796
## `Home Maker` -4.291e-01 1.591e-01 -2.698 0.006974 **
## Lawyer -2.943e-01 1.497e-01 -1.966 0.049341 *
## Manager -1.688e-01 1.341e-01 -1.258 0.208331
## Professional -2.919e-01 1.273e-01 -2.293 0.021851 *
## `z_Blue Collar` -1.532e-01 1.152e-01 -1.329 0.183733
## Pickup -8.215e-02 9.713e-02 -0.846 0.397662
## `Sports Car` 1.293e-01 1.073e-01 1.206 0.227912
## Van -8.559e-02 1.231e-01 -0.695 0.486781
## income_homeval_educ 7.887e-08 7.733e-08 1.020 0.307744
## sex_redcar_suv 4.606e-02 5.477e-02 0.841 0.400395
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5629.1 on 4878 degrees of freedom
## AIC: 5667.1
##
## Number of Fisher Scoring iterations: 4Van has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ KIDSDRIV + YOJ + TRAVTIME +
## CAR_USE + MVR_PTS + URBANICITY + Clerical + Doctor + `Home Maker` +
## Lawyer + Manager + Professional + `z_Blue Collar` + Pickup +
## `Sports Car` + income_homeval_educ + sex_redcar_suv, family = binomial(link = "logit"),
## data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.0345 -0.8032 -0.7546 1.5055 1.8724
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -9.391e-01 1.811e-01 -5.187 2.14e-07 ***
## KIDSDRIV 5.896e-02 6.287e-02 0.938 0.348302
## YOJ 8.796e-03 9.231e-03 0.953 0.340678
## TRAVTIME 2.459e-03 2.071e-03 1.187 0.235112
## CAR_USE -1.114e-01 8.967e-02 -1.242 0.214111
## MVR_PTS -1.311e-02 1.551e-02 -0.845 0.397869
## URBANICITY -9.971e-02 8.635e-02 -1.155 0.248176
## Clerical -4.670e-01 1.333e-01 -3.503 0.000459 ***
## Doctor -3.346e-01 2.256e-01 -1.483 0.138030
## `Home Maker` -4.302e-01 1.590e-01 -2.705 0.006839 **
## Lawyer -2.962e-01 1.497e-01 -1.979 0.047795 *
## Manager -1.708e-01 1.341e-01 -1.273 0.202877
## Professional -2.948e-01 1.272e-01 -2.318 0.020463 *
## `z_Blue Collar` -1.507e-01 1.152e-01 -1.309 0.190566
## Pickup -6.522e-02 9.409e-02 -0.693 0.488212
## `Sports Car` 1.404e-01 1.061e-01 1.323 0.185698
## income_homeval_educ 7.787e-08 7.735e-08 1.007 0.314084
## sex_redcar_suv 5.507e-02 5.329e-02 1.033 0.301440
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5629.6 on 4879 degrees of freedom
## AIC: 5665.6
##
## Number of Fisher Scoring iterations: 4Pickup has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ KIDSDRIV + YOJ + TRAVTIME +
## CAR_USE + MVR_PTS + URBANICITY + Clerical + Doctor + `Home Maker` +
## Lawyer + Manager + Professional + `z_Blue Collar` + `Sports Car` +
## income_homeval_educ + sex_redcar_suv, family = binomial(link = "logit"),
## data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.0354 -0.8034 -0.7548 1.5075 1.8581
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -9.611e-01 1.783e-01 -5.390 7.04e-08 ***
## KIDSDRIV 5.883e-02 6.286e-02 0.936 0.349373
## YOJ 8.808e-03 9.230e-03 0.954 0.339953
## TRAVTIME 2.487e-03 2.070e-03 1.201 0.229593
## CAR_USE -1.210e-01 8.853e-02 -1.367 0.171539
## MVR_PTS -1.340e-02 1.550e-02 -0.864 0.387392
## URBANICITY -9.961e-02 8.633e-02 -1.154 0.248581
## Clerical -4.695e-01 1.332e-01 -3.525 0.000424 ***
## Doctor -3.405e-01 2.255e-01 -1.510 0.131051
## `Home Maker` -4.319e-01 1.590e-01 -2.717 0.006590 **
## Lawyer -2.989e-01 1.496e-01 -1.998 0.045739 *
## Manager -1.730e-01 1.340e-01 -1.291 0.196795
## Professional -2.944e-01 1.272e-01 -2.315 0.020613 *
## `z_Blue Collar` -1.471e-01 1.150e-01 -1.279 0.200832
## `Sports Car` 1.529e-01 1.046e-01 1.461 0.143950
## income_homeval_educ 8.363e-08 7.686e-08 1.088 0.276552
## sex_redcar_suv 6.356e-02 5.195e-02 1.224 0.221103
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5630.0 on 4880 degrees of freedom
## AIC: 5664
##
## Number of Fisher Scoring iterations: 4MVR_PTS has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ KIDSDRIV + YOJ + TRAVTIME +
## CAR_USE + URBANICITY + Clerical + Doctor + `Home Maker` +
## Lawyer + Manager + Professional + `z_Blue Collar` + `Sports Car` +
## income_homeval_educ + sex_redcar_suv, family = binomial(link = "logit"),
## data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.0281 -0.8033 -0.7559 1.5076 1.8606
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -9.733e-01 1.777e-01 -5.476 4.34e-08 ***
## KIDSDRIV 5.584e-02 6.275e-02 0.890 0.373517
## YOJ 9.091e-03 9.221e-03 0.986 0.324207
## TRAVTIME 2.445e-03 2.070e-03 1.181 0.237452
## CAR_USE -1.256e-01 8.836e-02 -1.422 0.155097
## URBANICITY -1.131e-01 8.492e-02 -1.332 0.182828
## Clerical -4.693e-01 1.332e-01 -3.524 0.000426 ***
## Doctor -3.387e-01 2.254e-01 -1.502 0.132972
## `Home Maker` -4.288e-01 1.589e-01 -2.698 0.006973 **
## Lawyer -2.984e-01 1.496e-01 -1.995 0.046026 *
## Manager -1.669e-01 1.338e-01 -1.247 0.212345
## Professional -2.946e-01 1.272e-01 -2.316 0.020546 *
## `z_Blue Collar` -1.447e-01 1.150e-01 -1.258 0.208317
## `Sports Car` 1.479e-01 1.044e-01 1.416 0.156830
## income_homeval_educ 8.640e-08 7.676e-08 1.126 0.260344
## sex_redcar_suv 6.203e-02 5.191e-02 1.195 0.232076
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5630.8 on 4881 degrees of freedom
## AIC: 5662.8
##
## Number of Fisher Scoring iterations: 4KIDSDRIV has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ YOJ + TRAVTIME + CAR_USE +
## URBANICITY + Clerical + Doctor + `Home Maker` + Lawyer +
## Manager + Professional + `z_Blue Collar` + `Sports Car` +
## income_homeval_educ + sex_redcar_suv, family = binomial(link = "logit"),
## data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.0344 -0.8035 -0.7564 1.5106 1.8570
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -9.655e-01 1.775e-01 -5.438 5.38e-08 ***
## YOJ 9.535e-03 9.211e-03 1.035 0.300619
## TRAVTIME 2.454e-03 2.070e-03 1.186 0.235778
## CAR_USE -1.278e-01 8.831e-02 -1.447 0.147830
## URBANICITY -1.141e-01 8.491e-02 -1.344 0.179096
## Clerical -4.700e-01 1.332e-01 -3.530 0.000416 ***
## Doctor -3.480e-01 2.251e-01 -1.546 0.122167
## `Home Maker` -4.303e-01 1.589e-01 -2.708 0.006776 **
## Lawyer -3.040e-01 1.494e-01 -2.035 0.041890 *
## Manager -1.697e-01 1.338e-01 -1.269 0.204604
## Professional -2.992e-01 1.270e-01 -2.355 0.018520 *
## `z_Blue Collar` -1.438e-01 1.150e-01 -1.251 0.210998
## `Sports Car` 1.464e-01 1.044e-01 1.402 0.160977
## income_homeval_educ 8.546e-08 7.674e-08 1.114 0.265452
## sex_redcar_suv 6.307e-02 5.189e-02 1.216 0.224163
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5631.6 on 4882 degrees of freedom
## AIC: 5661.6
##
## Number of Fisher Scoring iterations: 4YOJ has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ TRAVTIME + CAR_USE + URBANICITY +
## Clerical + Doctor + `Home Maker` + Lawyer + Manager + Professional +
## `z_Blue Collar` + `Sports Car` + income_homeval_educ + sex_redcar_suv,
## family = binomial(link = "logit"), data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.9980 -0.8023 -0.7562 1.5101 1.8467
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -8.943e-01 1.633e-01 -5.476 4.36e-08 ***
## TRAVTIME 2.483e-03 2.070e-03 1.200 0.23028
## CAR_USE -1.289e-01 8.828e-02 -1.460 0.14440
## URBANICITY -1.109e-01 8.483e-02 -1.308 0.19099
## Clerical -4.352e-01 1.288e-01 -3.378 0.00073 ***
## Doctor -3.364e-01 2.249e-01 -1.496 0.13465
## `Home Maker` -4.472e-01 1.580e-01 -2.830 0.00466 **
## Lawyer -2.835e-01 1.481e-01 -1.913 0.05569 .
## Manager -1.469e-01 1.320e-01 -1.113 0.26585
## Professional -2.718e-01 1.243e-01 -2.187 0.02875 *
## `z_Blue Collar` -1.114e-01 1.106e-01 -1.007 0.31410
## `Sports Car` 1.439e-01 1.044e-01 1.378 0.16817
## income_homeval_educ 9.911e-08 7.557e-08 1.311 0.18970
## sex_redcar_suv 6.181e-02 5.186e-02 1.192 0.23328
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5632.6 on 4883 degrees of freedom
## AIC: 5660.6
##
## Number of Fisher Scoring iterations: 4Blue Collar has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ TRAVTIME + CAR_USE + URBANICITY +
## Clerical + Doctor + `Home Maker` + Lawyer + Manager + Professional +
## `Sports Car` + income_homeval_educ + sex_redcar_suv, family = binomial(link = "logit"),
## data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.9725 -0.8030 -0.7581 1.5168 1.8513
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -9.599e-01 1.499e-01 -6.402 1.53e-10 ***
## TRAVTIME 2.463e-03 2.070e-03 1.190 0.233999
## CAR_USE -1.410e-01 8.749e-02 -1.612 0.107042
## URBANICITY -1.098e-01 8.478e-02 -1.295 0.195222
## Clerical -3.683e-01 1.108e-01 -3.325 0.000884 ***
## Doctor -3.038e-01 2.229e-01 -1.363 0.172853
## `Home Maker` -3.832e-01 1.450e-01 -2.643 0.008211 **
## Lawyer -2.373e-01 1.412e-01 -1.681 0.092834 .
## Manager -9.445e-02 1.217e-01 -0.776 0.437707
## Professional -2.144e-01 1.108e-01 -1.935 0.053015 .
## `Sports Car` 1.432e-01 1.043e-01 1.373 0.169883
## income_homeval_educ 1.193e-07 7.290e-08 1.637 0.101648
## sex_redcar_suv 5.969e-02 5.179e-02 1.152 0.249118
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5633.7 on 4884 degrees of freedom
## AIC: 5659.7
##
## Number of Fisher Scoring iterations: 4sex_redcar_suv has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ TRAVTIME + CAR_USE + URBANICITY +
## Clerical + Doctor + `Home Maker` + Lawyer + Manager + Professional +
## `Sports Car` + income_homeval_educ, family = binomial(link = "logit"),
## data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.9737 -0.8025 -0.7608 1.5170 1.8408
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -8.766e-01 1.311e-01 -6.688 2.26e-11 ***
## TRAVTIME 2.452e-03 2.070e-03 1.185 0.236126
## CAR_USE -1.633e-01 8.524e-02 -1.916 0.055408 .
## URBANICITY -1.116e-01 8.476e-02 -1.317 0.187832
## Clerical -3.733e-01 1.106e-01 -3.375 0.000737 ***
## Doctor -3.054e-01 2.227e-01 -1.371 0.170311
## `Home Maker` -3.686e-01 1.444e-01 -2.553 0.010689 *
## Lawyer -2.434e-01 1.410e-01 -1.726 0.084261 .
## Manager -1.001e-01 1.215e-01 -0.824 0.409991
## Professional -2.215e-01 1.106e-01 -2.004 0.045102 *
## `Sports Car` 1.289e-01 1.036e-01 1.244 0.213418
## income_homeval_educ 1.145e-07 7.274e-08 1.574 0.115375
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5635.0 on 4885 degrees of freedom
## AIC: 5659
##
## Number of Fisher Scoring iterations: 4Manager has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ TRAVTIME + CAR_USE + URBANICITY +
## Clerical + Doctor + `Home Maker` + Lawyer + Professional +
## `Sports Car` + income_homeval_educ, family = binomial(link = "logit"),
## data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.9705 -0.8024 -0.7616 1.5177 1.8334
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -9.009e-01 1.278e-01 -7.047 1.83e-12 ***
## TRAVTIME 2.525e-03 2.068e-03 1.221 0.22206
## CAR_USE -1.366e-01 7.874e-02 -1.734 0.08286 .
## URBANICITY -1.219e-01 8.387e-02 -1.453 0.14609
## Clerical -3.469e-01 1.059e-01 -3.276 0.00105 **
## Doctor -2.422e-01 2.092e-01 -1.158 0.24687
## `Home Maker` -3.373e-01 1.394e-01 -2.420 0.01552 *
## Lawyer -1.931e-01 1.272e-01 -1.518 0.12893
## Professional -1.862e-01 1.019e-01 -1.827 0.06776 .
## `Sports Car` 1.300e-01 1.036e-01 1.255 0.20942
## income_homeval_educ 9.662e-08 6.962e-08 1.388 0.16519
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5635.7 on 4886 degrees of freedom
## AIC: 5657.7
##
## Number of Fisher Scoring iterations: 4Doctor has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ TRAVTIME + CAR_USE + URBANICITY +
## Clerical + `Home Maker` + Lawyer + Professional + `Sports Car` +
## income_homeval_educ, family = binomial(link = "logit"), data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.9652 -0.8035 -0.7636 1.5279 1.8248
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -9.079e-01 1.278e-01 -7.105 1.21e-12 ***
## TRAVTIME 2.539e-03 2.068e-03 1.228 0.21945
## CAR_USE -1.133e-01 7.622e-02 -1.487 0.13711
## URBANICITY -1.226e-01 8.387e-02 -1.462 0.14366
## Clerical -3.384e-01 1.057e-01 -3.202 0.00137 **
## `Home Maker` -3.223e-01 1.388e-01 -2.321 0.02027 *
## Lawyer -1.509e-01 1.220e-01 -1.237 0.21623
## Professional -1.633e-01 1.000e-01 -1.632 0.10270
## `Sports Car` 1.294e-01 1.036e-01 1.250 0.21145
## income_homeval_educ 6.385e-08 6.401e-08 0.998 0.31850
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5637.0 on 4887 degrees of freedom
## AIC: 5657
##
## Number of Fisher Scoring iterations: 4income_homeval_educ has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ TRAVTIME + CAR_USE + URBANICITY +
## Clerical + `Home Maker` + Lawyer + Professional + `Sports Car`,
## family = binomial(link = "logit"), data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.9741 -0.8023 -0.7656 1.5279 1.8242
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.874039 0.123063 -7.102 1.23e-12 ***
## TRAVTIME 0.002489 0.002066 1.204 0.228442
## CAR_USE -0.115902 0.076171 -1.522 0.128106
## URBANICITY -0.110154 0.082884 -1.329 0.183846
## Clerical -0.366179 0.101823 -3.596 0.000323 ***
## `Home Maker` -0.342347 0.137301 -2.493 0.012653 *
## Lawyer -0.126110 0.119493 -1.055 0.291256
## Professional -0.162724 0.100042 -1.627 0.103830
## `Sports Car` 0.123735 0.103394 1.197 0.231410
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5638.0 on 4888 degrees of freedom
## AIC: 5656
##
## Number of Fisher Scoring iterations: 4Lawyer has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ TRAVTIME + CAR_USE + URBANICITY +
## Clerical + `Home Maker` + Professional + `Sports Car`, family = binomial(link = "logit"),
## data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.9632 -0.8047 -0.7658 1.5350 1.8145
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.904444 0.119743 -7.553 4.25e-14 ***
## TRAVTIME 0.002484 0.002066 1.202 0.229291
## CAR_USE -0.087183 0.071255 -1.224 0.221131
## URBANICITY -0.114165 0.082798 -1.379 0.167944
## Clerical -0.338605 0.098482 -3.438 0.000585 ***
## `Home Maker` -0.311776 0.134289 -2.322 0.020250 *
## Professional -0.136250 0.096903 -1.406 0.159713
## `Sports Car` 0.126405 0.103357 1.223 0.221330
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5639.1 on 4889 degrees of freedom
## AIC: 5655.1
##
## Number of Fisher Scoring iterations: 4TRAVTIME has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ CAR_USE + URBANICITY + Clerical +
## `Home Maker` + Professional + `Sports Car`, family = binomial(link = "logit"),
## data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.9050 -0.8118 -0.7665 1.5346 1.7838
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.80917 0.08947 -9.044 < 2e-16 ***
## CAR_USE -0.08358 0.07118 -1.174 0.240259
## URBANICITY -0.13172 0.08148 -1.617 0.105957
## Clerical -0.33864 0.09847 -3.439 0.000584 ***
## `Home Maker` -0.30812 0.13423 -2.295 0.021711 *
## Professional -0.13365 0.09687 -1.380 0.167657
## `Sports Car` 0.12820 0.10333 1.241 0.214694
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5640.6 on 4890 degrees of freedom
## AIC: 5654.6
##
## Number of Fisher Scoring iterations: 4CAR_USE has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ URBANICITY + Clerical + `Home Maker` +
## Professional + `Sports Car`, family = binomial(link = "logit"),
## data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.8927 -0.7980 -0.7595 1.5474 1.7528
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.85766 0.07949 -10.790 < 2e-16 ***
## URBANICITY -0.12330 0.08113 -1.520 0.12858
## Clerical -0.31278 0.09596 -3.259 0.00112 **
## `Home Maker` -0.27429 0.13112 -2.092 0.03646 *
## Professional -0.11468 0.09550 -1.201 0.22982
## `Sports Car` 0.14329 0.10254 1.397 0.16229
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5642.0 on 4891 degrees of freedom
## AIC: 5654
##
## Number of Fisher Scoring iterations: 4Professional has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ URBANICITY + Clerical + `Home Maker` +
## `Sports Car`, family = binomial(link = "logit"), data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.8849 -0.7909 -0.7909 1.5564 1.7526
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.87977 0.07740 -11.367 < 2e-16 ***
## URBANICITY -0.12210 0.08111 -1.505 0.13226
## Clerical -0.29154 0.09438 -3.089 0.00201 **
## `Home Maker` -0.25318 0.12999 -1.948 0.05145 .
## `Sports Car` 0.14427 0.10252 1.407 0.15937
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5643.4 on 4892 degrees of freedom
## AIC: 5653.4
##
## Number of Fisher Scoring iterations: 4Sports Car has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ URBANICITY + Clerical + `Home Maker`,
## family = binomial(link = "logit"), data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.8378 -0.7957 -0.7957 1.5604 1.7443
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.86640 0.07678 -11.285 < 2e-16 ***
## URBANICITY -0.12129 0.08110 -1.496 0.13476
## Clerical -0.28720 0.09431 -3.045 0.00232 **
## `Home Maker` -0.23226 0.12906 -1.800 0.07193 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5645.4 on 4893 degrees of freedom
## AIC: 5653.4
##
## Number of Fisher Scoring iterations: 4URBANICITY has the (highest p value) lowest affect on the target and will be removed next.
##
## Call:
## glm(formula = train1$TARGET_FLAG ~ Clerical + `Home Maker`, family = binomial(link = "logit"),
## data = train3)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.8025 -0.8025 -0.8025 1.6061 1.7242
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.96776 0.03660 -26.440 < 2e-16 ***
## Clerical -0.26228 0.09273 -2.828 0.00468 **
## `Home Maker` -0.20824 0.12798 -1.627 0.10372
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5657.6 on 4896 degrees of freedom
## Residual deviance: 5647.6 on 4894 degrees of freedom
## AIC: 5653.6
##
## Number of Fisher Scoring iterations: 4Home Maker has the (highest p value) lowest affect on the target and will be removed next.
## (Intercept) Clerical
## -0.19144339 -0.04747191The variable that is likely to result in a customer not having a claim is: having a clerical job. This seems like a very unlikely model to use to predict the likelihood a customer would get into an accident.
Prediction from Model 3
## [1] 0.23## [1] 0.4842399The cutoff associated with a point the farthest distance from the ROC curve is 0.23. I will use 0.23 as the cutoff for making predictions. A value above 0.23 will be assigned a target of one and a value below 0.23 wil be assigned a value of zero. The area under the curve is 0.48.
Confusion Matrix
## true
## pred 0 1
## 0 337 147
## 1 2070 710The model predicted 337 0’s that were actually 0. The model predicted 147 0’s that were actually 1.
The model predicted 2,070 1’s that were actaully 0. The model predicted 710 1’s that were actually 1.
## accuracy error precision sensitivity specificity f1
## 1 0.3207721 0.6792279 0.1165687 0.1400083 0.8284714 0.1272178Building a Multiple Linear Regression Model to Predict the Cost of a Claim
I will start with logit model 2, since it has the highest accuracy and the highest area under the ROC curve. I will use model 2 to build a muliple linear regression model to predict the claim amount. A customer who is predicted to not have a claim (TARGET_FLAG=0) will be predicted to have 0 dollars for the TARGET_AMT.
##
## Call:
## lm(formula = train2a$TARGET_AMT ~ ., data = train2a)
##
## Residuals:
## Min 1Q Median 3Q Max
## -5217 -1713 -726 397 76398
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -3.236e+02 3.101e+02 -1.043 0.296869
## KIDSDRIV 3.129e+02 1.272e+02 2.459 0.013955 *
## INCOME -5.603e-03 1.759e-03 -3.186 0.001453 **
## PARENT1 9.679e+02 2.167e+02 4.467 8.11e-06 ***
## MSTATUS -5.577e+02 1.464e+02 -3.810 0.000141 ***
## EDUCATION -1.418e+02 6.641e+01 -2.135 0.032828 *
## TRAVTIME 1.363e+01 3.946e+00 3.454 0.000558 ***
## CAR_USE 7.930e+02 1.564e+02 5.070 4.13e-07 ***
## BLUEBOOK 7.132e-03 9.537e-03 0.748 0.454566
## TIF -4.742e+01 1.493e+01 -3.175 0.001506 **
## CLM_FREQ 8.311e+01 6.040e+01 1.376 0.168879
## REVOKED 5.237e+02 1.924e+02 2.722 0.006510 **
## MVR_PTS 1.307e+02 3.252e+01 4.020 5.90e-05 ***
## URBANICITY 1.854e+03 1.688e+02 10.981 < 2e-16 ***
## Manager -9.728e+02 2.040e+02 -4.768 1.92e-06 ***
## Professional -3.946e+02 1.857e+02 -2.125 0.033635 *
## `Panel Truck` 7.261e+02 3.045e+02 2.385 0.017138 *
## Pickup 3.023e+02 2.015e+02 1.500 0.133597
## `Sports Car` 5.841e+02 2.293e+02 2.548 0.010866 *
## Van 1.570e+02 2.491e+02 0.630 0.528550
## z_SUV 4.965e+02 1.711e+02 2.901 0.003735 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 4322 on 4876 degrees of freedom
## Multiple R-squared: 0.08062, Adjusted R-squared: 0.07685
## F-statistic: 21.38 on 20 and 4876 DF, p-value: < 2.2e-16Van has the lowest affect on claim amount and will be removed.
##
## Call:
## lm(formula = train2a$TARGET_AMT ~ KIDSDRIV + INCOME + PARENT1 +
## MSTATUS + EDUCATION + TRAVTIME + CAR_USE + BLUEBOOK + TIF +
## CLM_FREQ + REVOKED + MVR_PTS + URBANICITY + Manager + Professional +
## `Panel Truck` + Pickup + `Sports Car` + z_SUV, data = train2a)
##
## Residuals:
## Min 1Q Median 3Q Max
## -5206 -1714 -718 385 76344
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -3.111e+02 3.095e+02 -1.005 0.314846
## KIDSDRIV 3.132e+02 1.272e+02 2.462 0.013840 *
## INCOME -5.582e-03 1.758e-03 -3.175 0.001510 **
## PARENT1 9.623e+02 2.165e+02 4.445 8.98e-06 ***
## MSTATUS -5.590e+02 1.464e+02 -3.820 0.000135 ***
## EDUCATION -1.410e+02 6.640e+01 -2.124 0.033744 *
## TRAVTIME 1.363e+01 3.945e+00 3.454 0.000556 ***
## CAR_USE 8.174e+02 1.515e+02 5.394 7.21e-08 ***
## BLUEBOOK 7.952e-03 9.447e-03 0.842 0.399953
## TIF -4.721e+01 1.493e+01 -3.162 0.001575 **
## CLM_FREQ 8.377e+01 6.039e+01 1.387 0.165428
## REVOKED 5.264e+02 1.923e+02 2.737 0.006226 **
## MVR_PTS 1.313e+02 3.251e+01 4.038 5.48e-05 ***
## URBANICITY 1.854e+03 1.688e+02 10.983 < 2e-16 ***
## Manager -9.688e+02 2.039e+02 -4.751 2.08e-06 ***
## Professional -3.897e+02 1.855e+02 -2.101 0.035729 *
## `Panel Truck` 6.577e+02 2.845e+02 2.312 0.020827 *
## Pickup 2.591e+02 1.894e+02 1.367 0.171542
## `Sports Car` 5.522e+02 2.236e+02 2.470 0.013548 *
## z_SUV 4.642e+02 1.633e+02 2.843 0.004490 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 4321 on 4877 degrees of freedom
## Multiple R-squared: 0.08054, Adjusted R-squared: 0.07696
## F-statistic: 22.49 on 19 and 4877 DF, p-value: < 2.2e-16BLUEBOOK has the lowest affect on claim amount and will be removed.
##
## Call:
## lm(formula = train2a$TARGET_AMT ~ KIDSDRIV + INCOME + PARENT1 +
## MSTATUS + EDUCATION + TRAVTIME + CAR_USE + TIF + CLM_FREQ +
## REVOKED + MVR_PTS + URBANICITY + Manager + Professional +
## `Panel Truck` + Pickup + `Sports Car` + z_SUV, data = train2a)
##
## Residuals:
## Min 1Q Median 3Q Max
## -5209 -1719 -723 382 76375
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -2.084e+02 2.844e+02 -0.733 0.463712
## KIDSDRIV 3.160e+02 1.272e+02 2.485 0.012982 *
## INCOME -5.215e-03 1.703e-03 -3.062 0.002214 **
## PARENT1 9.617e+02 2.165e+02 4.442 9.09e-06 ***
## MSTATUS -5.577e+02 1.463e+02 -3.811 0.000140 ***
## EDUCATION -1.380e+02 6.630e+01 -2.082 0.037437 *
## TRAVTIME 1.363e+01 3.945e+00 3.456 0.000553 ***
## CAR_USE 8.257e+02 1.512e+02 5.461 4.97e-08 ***
## TIF -4.727e+01 1.493e+01 -3.167 0.001552 **
## CLM_FREQ 8.231e+01 6.036e+01 1.364 0.172746
## REVOKED 5.261e+02 1.923e+02 2.735 0.006253 **
## MVR_PTS 1.315e+02 3.251e+01 4.044 5.33e-05 ***
## URBANICITY 1.854e+03 1.688e+02 10.987 < 2e-16 ***
## Manager -9.666e+02 2.039e+02 -4.740 2.19e-06 ***
## Professional -3.834e+02 1.854e+02 -2.068 0.038688 *
## `Panel Truck` 7.375e+02 2.683e+02 2.749 0.005997 **
## Pickup 2.282e+02 1.859e+02 1.228 0.219566
## `Sports Car` 5.206e+02 2.204e+02 2.362 0.018217 *
## z_SUV 4.330e+02 1.590e+02 2.723 0.006495 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 4321 on 4878 degrees of freedom
## Multiple R-squared: 0.08041, Adjusted R-squared: 0.07702
## F-statistic: 23.7 on 18 and 4878 DF, p-value: < 2.2e-16Pickup has the lowest affect on claim amount and will be removed.
##
## Call:
## lm(formula = train2a$TARGET_AMT ~ KIDSDRIV + INCOME + PARENT1 +
## MSTATUS + EDUCATION + TRAVTIME + CAR_USE + TIF + CLM_FREQ +
## REVOKED + MVR_PTS + URBANICITY + Manager + Professional +
## `Panel Truck` + `Sports Car` + z_SUV, data = train2a)
##
## Residuals:
## Min 1Q Median 3Q Max
## -5208 -1715 -727 372 76258
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -1.410e+02 2.791e+02 -0.505 0.613393
## KIDSDRIV 3.163e+02 1.272e+02 2.487 0.012919 *
## INCOME -5.381e-03 1.698e-03 -3.169 0.001537 **
## PARENT1 9.571e+02 2.165e+02 4.422 1.00e-05 ***
## MSTATUS -5.564e+02 1.463e+02 -3.802 0.000145 ***
## EDUCATION -1.394e+02 6.629e+01 -2.102 0.035569 *
## TRAVTIME 1.356e+01 3.945e+00 3.438 0.000591 ***
## CAR_USE 8.699e+02 1.469e+02 5.923 3.38e-09 ***
## TIF -4.693e+01 1.493e+01 -3.144 0.001675 **
## CLM_FREQ 8.238e+01 6.036e+01 1.365 0.172395
## REVOKED 5.334e+02 1.922e+02 2.775 0.005549 **
## MVR_PTS 1.323e+02 3.250e+01 4.071 4.76e-05 ***
## URBANICITY 1.854e+03 1.688e+02 10.987 < 2e-16 ***
## Manager -9.580e+02 2.038e+02 -4.701 2.66e-06 ***
## Professional -3.787e+02 1.853e+02 -2.043 0.041069 *
## `Panel Truck` 6.410e+02 2.565e+02 2.499 0.012487 *
## `Sports Car` 4.523e+02 2.133e+02 2.121 0.033996 *
## z_SUV 3.645e+02 1.489e+02 2.448 0.014419 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 4321 on 4879 degrees of freedom
## Multiple R-squared: 0.08013, Adjusted R-squared: 0.07692
## F-statistic: 25 on 17 and 4879 DF, p-value: < 2.2e-16CLM_FREQ has the lowest affect on claim amount and will be removed.
##
## Call:
## lm(formula = train2a$TARGET_AMT ~ KIDSDRIV + INCOME + PARENT1 +
## MSTATUS + EDUCATION + TRAVTIME + CAR_USE + TIF + REVOKED +
## MVR_PTS + URBANICITY + Manager + Professional + `Panel Truck` +
## `Sports Car` + z_SUV, data = train2a)
##
## Residuals:
## Min 1Q Median 3Q Max
## -5317 -1714 -737 369 76377
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -1.451e+02 2.791e+02 -0.520 0.603117
## KIDSDRIV 3.246e+02 1.270e+02 2.555 0.010649 *
## INCOME -5.514e-03 1.695e-03 -3.253 0.001151 **
## PARENT1 9.513e+02 2.164e+02 4.395 1.13e-05 ***
## MSTATUS -5.688e+02 1.461e+02 -3.894 9.98e-05 ***
## EDUCATION -1.392e+02 6.630e+01 -2.099 0.035827 *
## TRAVTIME 1.383e+01 3.941e+00 3.508 0.000455 ***
## CAR_USE 8.809e+02 1.467e+02 6.006 2.04e-09 ***
## TIF -4.702e+01 1.493e+01 -3.150 0.001641 **
## REVOKED 5.349e+02 1.923e+02 2.782 0.005421 **
## MVR_PTS 1.483e+02 3.032e+01 4.892 1.03e-06 ***
## URBANICITY 1.906e+03 1.645e+02 11.584 < 2e-16 ***
## Manager -9.604e+02 2.038e+02 -4.712 2.52e-06 ***
## Professional -3.815e+02 1.854e+02 -2.058 0.039610 *
## `Panel Truck` 6.499e+02 2.564e+02 2.534 0.011300 *
## `Sports Car` 4.706e+02 2.129e+02 2.211 0.027096 *
## z_SUV 3.708e+02 1.488e+02 2.491 0.012765 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 4322 on 4880 degrees of freedom
## Multiple R-squared: 0.07978, Adjusted R-squared: 0.07676
## F-statistic: 26.44 on 16 and 4880 DF, p-value: < 2.2e-16All of the remaining variables have p values lower that 0.05, which indicates that they are signficant. A p value below 0.05 indicates strong evidence against the null hypothesis that the variable in question does not affect the claim amount.
The adjusted R squared value is .076. 7.6% of the variability in claim amount is accounted for by the model. This is a very unimpressive model.
The F statistic is 26.44, which is high, and indicates that these variables are signficiant.
The following graphs look for patterns in the residuals.
There are some residuals whose values are very high, indicating a large deviation from the model. The residuals follow the line on the QQ plot until a point, and then deviate greatly. This is because some claims are for very large amounts and the model does not predict those very high values.
Prediction from Model 2a
The root mean square error from model 2a is
## [1] 4907.037The cost of claims predicted by model 2a is off, on average, by $4907. This is an exorbitant amount to be off by on average. The claims for very large amounts are not predicted by my model and are causing the error to be high.
Building a Logarthmic model - Using backward elimination
To try to better account for the very high claim amounts, I will try a logarithmic model. This will also be built off of logit model 2.
##
## Call:
## lm(formula = log(train2b$TARGET_AMT) ~ ., data = train2b)
##
## Residuals:
## Min 1Q Median 3Q Max
## -19.836 -6.368 -2.433 6.303 27.055
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -1.339e+01 6.178e-01 -21.670 < 2e-16 ***
## KIDSDRIV 1.271e+00 2.534e-01 5.015 5.48e-07 ***
## INCOME -1.826e-05 3.504e-06 -5.211 1.96e-07 ***
## PARENT1 2.025e+00 4.316e-01 4.690 2.80e-06 ***
## MSTATUS -1.878e+00 2.916e-01 -6.439 1.32e-10 ***
## EDUCATION -5.609e-01 1.323e-01 -4.240 2.27e-05 ***
## TRAVTIME 4.856e-02 7.860e-03 6.178 7.00e-10 ***
## CAR_USE 2.779e+00 3.116e-01 8.920 < 2e-16 ***
## BLUEBOOK -7.314e-05 1.900e-05 -3.850 0.00012 ***
## TIF -1.778e-01 2.975e-02 -5.976 2.45e-09 ***
## CLM_FREQ 5.613e-01 1.203e-01 4.665 3.17e-06 ***
## REVOKED 3.031e+00 3.832e-01 7.910 3.16e-15 ***
## MVR_PTS 4.441e-01 6.478e-02 6.856 7.97e-12 ***
## URBANICITY 6.518e+00 3.362e-01 19.384 < 2e-16 ***
## Manager -3.063e+00 4.064e-01 -7.537 5.69e-14 ***
## Professional -8.244e-01 3.699e-01 -2.228 0.02590 *
## `Panel Truck` 1.761e+00 6.066e-01 2.904 0.00370 **
## Pickup 1.209e+00 4.014e-01 3.011 0.00262 **
## `Sports Car` 2.825e+00 4.567e-01 6.187 6.65e-10 ***
## Van 1.579e+00 4.962e-01 3.182 0.00147 **
## z_SUV 2.191e+00 3.409e-01 6.426 1.44e-10 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 8.609 on 4876 degrees of freedom
## Multiple R-squared: 0.2246, Adjusted R-squared: 0.2214
## F-statistic: 70.63 on 20 and 4876 DF, p-value: < 2.2e-16All of the remaining variables have p values lower that 0.05, which indicates that they are signficant. A p value below 0.05 indicates strong evidence against the null hypothesis that the variable in question does not affect the claim amount.
The adjusted R squared value is .22. 22% of the variability in claim amount is accounted for by the model. This is a significant improvement over the last model.
The F statistic is 70.63, which is high, and indicates that these variables are signficiant.
The following graphs look for patterns in the residuals.
The residuals show a pattern and the residuals deviate from the line on the QQ plot. These indicate that this is not a good model.
Prediction from Model 2b- Logarithmic Model
The root mean square error from model 2 is
## [1] 5219.295The cost of claims predicted by model 2b is off, on average, by $5219. This is even higher amount to be off by than the previous model. The claims for very large amounts are not predicted by my model and are causing the error to be high.
Model Selection
In order to predict the TARGET_FLG, which is a binary variable that describes whether or not a customer has a claim, I wil use model 2. This is a logit model that was built using backward elimination starting with all of the variables.
When comparing the three models, the accuracy is greatest for the 2nd model. The precision is also highest for the second model. (The precision is the ratio of correct predictions of zero to total predictions of zero.) The sensitivity is highest for model 2. (The sensitivity is the ratio of the correct predictions of zero to all cases in which the target is zero.) The specificity is the greatest for model 3. (The specificity is the ratio of the correct predictions of 1 to all cases in which the target is one.) The F1 score is highest for the second model. (The F1 score is equal to 2xPrecisionxSensitivity/(Precision+Sensitivity) and gives a balance between the precision and sensitivity.) The area under the roc curve is the farthest from 0.5 for model 2. (The farther the area is from 0.5, the better the model.)
I will use model 2 to make a prediction for the test data because it has a higher accuracy, precision, sensitivity, F1 score and area under the ROC curve than the other models.
The predictions for the evaluation set are below:
## 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
## 0 0 0 0 0 0 1 1 0 0 0 1 1 0 0
## 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
## 1 1 0 1 1 0 1 0 1 1 1 1 1 0 0
## 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45
## 0 1 0 0 0 0 0 0 0 1 0 1 0 1 0
## 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
## 0 0 1 0 1 0 0 1 0 0 0 0 1 0 1
## 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75
## 0 0 1 0 0 1 1 1 0 0 0 0 1 1 1
## 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90
## 0 1 0 0 0 1 1 1 0 1 1 0 0 0 1
## 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105
## 0 0 0 0 0 0 0 0 1 0 1 1 1 1 0
## 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120
## 0 0 0 1 0 1 0 0 0 1 0 0 1 1 0
## 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135
## 0 1 1 1 0 1 1 0 0 0 0 0 0 0 0
## 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150
## 0 1 1 0 0 0 1 0 0 0 1 0 0 0 0
## 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165
## 1 1 1 1 0 1 0 0 1 1 0 0 0 0 1
## 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180
## 0 0 0 1 0 0 1 0 1 1 1 1 1 1 1
## 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195
## 1 0 0 0 0 0 0 0 0 0 1 1 1 0 0
## 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210
## 1 1 0 0 0 0 1 0 0 0 0 1 0 0 0
## 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225
## 0 0 1 1 0 0 1 0 0 0 0 0 1 1 0
## 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240
## 1 1 1 1 0 0 0 1 0 0 0 0 0 0 1
## 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255
## 0 0 1 0 0 1 1 1 0 1 1 1 0 1 1
## 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270
## 1 0 0 1 0 0 0 0 0 0 0 0 0 1 1
## 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285
## 1 0 0 1 0 0 1 0 0 0 0 0 0 1 1
## 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300
## 1 0 1 1 1 1 0 0 1 0 1 0 1 0 0
## 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315
## 0 0 0 1 1 1 0 1 0 0 1 0 0 1 0
## 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330
## 0 0 0 1 0 0 1 0 0 1 1 1 0 1 0
## 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345
## 0 0 1 0 0 0 0 1 0 0 1 1 1 1 0
## 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360
## 0 0 0 0 0 0 0 1 1 0 1 1 0 0 0
## 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375
## 1 0 0 1 0 1 0 1 0 0 0 0 1 0 0
## 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390
## 1 0 0 0 0 0 1 0 0 0 1 1 0 0 1
## 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405
## 0 0 0 0 0 1 0 1 0 1 1 0 0 0 0
## 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420
## 0 0 1 0 0 0 1 0 0 1 1 0 1 0 0
## 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435
## 1 1 1 0 0 1 0 0 1 1 1 0 0 0 1
## 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450
## 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1
## 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465
## 0 0 1 0 1 1 1 1 0 1 0 0 0 1 0
## 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480
## 0 1 1 0 0 0 1 0 0 0 0 1 1 0 1
## 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495
## 0 0 0 1 1 1 0 1 0 1 1 0 0 0 0
## 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510
## 1 0 0 1 0 0 0 1 0 1 0 1 0 0 0
## 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525
## 0 0 0 0 0 0 1 1 0 1 0 0 0 0 0
## 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540
## 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0
## 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555
## 0 0 1 1 0 0 0 1 1 0 0 0 0 1 0
## 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570
## 0 1 0 1 0 0 1 0 0 1 0 1 1 1 1
## 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585
## 0 0 0 0 0 0 0 0 1 0 0 1 0 1 0
## 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600
## 0 0 0 1 1 1 0 0 1 1 1 1 0 0 1
## 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615
## 1 0 0 0 1 1 1 0 0 0 0 1 0 0 1
## 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630
## 0 0 1 0 1 0 0 0 0 0 1 1 0 0 1
## 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645
## 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0
## 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660
## 1 1 0 0 0 0 0 1 0 0 0 1 0 0 0
## 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675
## 1 1 1 0 0 1 1 0 0 0 0 1 1 0 1
## 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690
## 0 0 1 0 0 1 0 1 0 0 0 0 0 0 0
## 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705
## 0 0 0 0 0 1 0 1 0 0 0 0 1 0 0
## 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720
## 0 1 1 0 0 0 0 1 0 0 0 0 0 0 0
## 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735
## 1 0 0 0 0 0 0 0 0 0 1 1 0 0 1
## 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750
## 1 0 0 0 0 1 1 1 0 0 1 1 0 0 0
## 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765
## 0 0 1 1 0 1 1 0 0 1 0 1 1 1 1
## 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780
## 1 1 0 1 0 1 1 0 1 0 1 0 1 0 0
## 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795
## 0 1 0 0 0 1 0 0 0 0 0 0 1 0 0
## 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810
## 1 0 1 1 1 0 0 0 1 0 0 0 0 0 1
## 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825
## 1 0 0 0 0 0 0 1 1 0 1 0 1 1 1
## 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840
## 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0
## 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855
## 0 0 0 0 0 0 1 1 1 1 1 0 0 0 1
## 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870
## 0 0 0 1 0 0 1 0 0 0 0 1 0 0 1
## 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885
## 0 1 0 1 1 0 0 0 0 1 1 0 0 0 1
## 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900
## 1 1 0 0 0 0 0 0 0 0 1 0 0 1 1
## 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915
## 0 0 1 0 0 0 1 0 1 1 1 0 0 0 0
## 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930
## 1 1 1 0 1 0 0 0 0 0 0 0 0 1 0
## 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945
## 0 1 0 1 0 1 0 1 1 1 1 0 1 0 1
## 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960
## 0 0 0 0 1 0 0 0 1 0 0 0 0 1 0
## 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975
## 1 1 1 0 0 1 1 0 0 0 1 0 0 1 0
## 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990
## 1 1 0 0 1 0 0 1 1 1 1 1 0 0 1
## 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005
## 0 0 1 0 1 0 0 0 0 0 1 1 1 0 1
## 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020
## 0 0 1 1 0 1 0 0 0 0 1 0 1 1 0
## 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035
## 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0
## 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050
## 0 0 0 0 0 0 1 1 0 1 0 1 1 1 1
## 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065
## 1 1 1 0 1 0 0 0 1 1 0 1 1 0 0
## 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080
## 0 0 0 0 1 0 0 0 1 0 0 0 0 1 0
## 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095
## 1 1 0 0 1 1 0 0 0 0 0 0 0 1 0
## 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110
## 0 0 0 1 1 0 0 1 1 1 1 0 0 0 1
## 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125
## 1 1 0 0 0 0 1 1 1 0 1 1 0 0 0
## 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140
## 0 1 0 0 0 0 0 1 0 1 0 0 0 0 0
## 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155
## 0 0 0 1 1 0 0 1 0 1 1 1 0 1 1
## 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170
## 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0
## 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185
## 1 1 1 1 0 0 1 0 0 1 0 1 0 1 1
## 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200
## 0 0 0 0 0 1 0 0 1 0 0 0 0 1 0
## 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215
## 0 0 0 0 0 0 1 0 0 1 0 0 1 0 0
## 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230
## 0 1 1 0 0 0 0 1 0 1 0 0 0 0 1
## 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245
## 0 0 1 1 0 0 0 1 0 0 1 0 0 0 0
## 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260
## 1 1 0 0 0 0 1 0 1 1 0 1 0 1 0
## 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275
## 0 0 0 1 0 0 1 0 1 0 0 0 0 0 0
## 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290
## 0 1 0 0 0 1 0 0 1 1 0 1 0 0 0
## 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305
## 1 0 0 1 1 0 1 0 1 1 0 0 0 0 0
## 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320
## 0 1 1 0 1 1 0 1 0 0 0 0 0 0 1
## 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335
## 0 1 1 0 0 1 0 1 0 0 0 0 0 0 1
## 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350
## 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0
## 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365
## 0 1 0 0 0 0 0 1 0 0 0 0 1 0 0
## 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380
## 0 1 1 1 1 0 0 0 0 0 1 1 0 0 0
## 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395
## 1 1 1 0 0 0 0 0 0 1 1 1 0 1 0
## 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410
## 0 0 1 1 0 0 0 1 1 1 0 0 0 0 1
## 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425
## 0 0 0 0 0 0 0 0 1 0 0 1 1 0 1
## 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440
## 1 0 0 1 0 0 0 0 0 0 1 1 1 0 0
## 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455
## 0 1 1 0 0 1 1 0 0 0 0 0 1 0 0
## 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470
## 0 1 0 0 0 0 0 0 1 0 1 0 0 0 1
## 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485
## 0 0 0 1 1 0 1 0 0 0 1 0 1 0 1
## 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500
## 0 0 0 1 0 0 1 0 0 1 1 0 0 0 0
## 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515
## 0 1 0 1 0 0 1 0 0 0 0 0 0 0 1
## 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530
## 1 0 1 0 0 1 0 0 0 0 1 0 1 1 1
## 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545
## 1 1 0 0 1 0 0 1 1 1 0 0 0 1 1
## 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560
## 0 0 0 0 0 0 0 0 1 0 0 0 0 1 1
## 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575
## 0 1 0 1 1 0 0 0 0 1 1 0 0 1 0
## 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590
## 1 1 0 0 0 0 0 0 0 0 0 0 1 0 1
## 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605
## 1 1 0 0 0 0 0 0 0 1 1 0 1 1 0
## 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620
## 1 0 0 0 1 1 0 0 1 1 1 0 1 0 1
## 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635
## 1 0 1 0 0 0 0 0 0 1 0 1 0 1 0
## 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650
## 0 1 0 0 0 0 0 0 0 1 0 0 0 0 1
## 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665
## 0 0 0 0 1 1 0 0 0 0 1 1 1 1 1
## 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680
## 0 1 0 1 0 1 0 1 0 0 0 0 0 0 0
## 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695
## 0 1 1 0 1 0 0 1 0 0 0 0 0 1 1
## 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710
## 1 0 1 1 0 1 0 0 1 0 0 1 0 1 0
## 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725
## 0 0 1 0 1 0 0 0 0 0 0 1 0 1 1
## 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740
## 0 0 0 1 1 0 0 0 0 0 0 1 0 0 0
## 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755
## 1 0 0 0 1 0 0 0 1 0 0 0 0 1 0
## 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770
## 0 0 1 0 0 1 0 0 0 1 1 0 0 0 0
## 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785
## 0 0 0 1 1 0 1 0 1 0 0 0 0 1 0
## 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800
## 0 0 0 1 1 0 0 0 0 0 1 0 0 0 1
## 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815
## 1 0 0 0 0 1 1 1 0 0 0 0 1 0 1
## 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830
## 0 0 0 0 0 0 0 0 1 0 1 0 1 0 0
## 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845
## 0 0 0 0 1 0 0 1 0 1 0 0 0 0 0
## 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860
## 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0
## 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875
## 1 0 0 0 0 0 0 0 1 1 1 0 0 0 1
## 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890
## 0 0 1 1 1 0 0 0 0 1 0 0 0 1 0
## 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905
## 0 0 0 0 1 0 0 0 0 1 0 1 0 1 0
## 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920
## 0 0 0 1 1 0 1 0 0 1 0 0 0 0 0
## 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935
## 0 1 0 0 0 0 0 0 0 0 0 0 1 1 0
## 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950
## 1 0 0 0 0 0 0 0 0 0 0 1 0 1 0
## 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965
## 0 0 1 1 0 0 0 0 0 1 1 1 1 0 0
## 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980
## 0 0 1 0 0 0 0 1 1 0 0 0 1 1 0
## 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995
## 0 1 0 0 0 0 0 0 0 1 0 1 1 1 0
## 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
## 0 1 1 0 0 1 1 1 0 1 0 1 0 1 1
## 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025
## 1 0 1 0 0 1 0 1 1 0 0 0 1 0 0
## 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040
## 0 0 0 0 1 0 0 0 0 1 1 0 0 1 0
## 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055
## 0 0 0 1 0 0 0 1 0 0 1 0 1 0 1
## 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070
## 0 0 0 0 0 0 1 1 0 0 0 1 1 0 0
## 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085
## 0 0 1 1 0 0 1 0 0 1 0 0 1 0 0
## 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100
## 0 1 1 0 1 0 0 1 0 1 1 0 1 1 1
## 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115
## 1 1 1 0 0 0 1 0 0 0 1 0 1 0 0
## 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130
## 0 1 0 1 0 0 0 1 1 0 0 1 0 0 0
## 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141
## 0 0 0 1 1 0 0 0 0 0 0I will use model 2a to predict the claim amount because that model had the lower root mean square error.
## 1 2 3 4 5 6 7 8
## 0.000 0.000 0.000 0.000 0.000 0.000 2369.262 3466.281
## 9 10 11 12 13 14 15 16
## 0.000 0.000 0.000 3799.551 4620.172 0.000 0.000 3737.505
## 17 18 19 20 21 22 23 24
## 3431.916 0.000 2730.605 2290.083 0.000 1832.988 0.000 2081.918
## 25 26 27 28 29 30 31 32
## 2041.693 2304.514 1851.031 3137.272 0.000 0.000 0.000 2082.547
## 33 34 35 36 37 38 39 40
## 0.000 0.000 0.000 0.000 0.000 0.000 0.000 2964.835
## 41 42 43 44 45 46 47 48
## 0.000 2855.948 0.000 2714.003 0.000 0.000 0.000 1905.791
## 49 50 51 52 53 54 55 56
## 0.000 3244.206 0.000 0.000 3395.802 0.000 0.000 0.000
## 57 58 59 60 61 62 63 64
## 0.000 2499.167 0.000 2547.963 0.000 0.000 1979.590 0.000
## 65 66 67 68 69 70 71 72
## 0.000 1467.866 4016.118 3949.387 0.000 0.000 0.000 0.000
## 73 74 75 76 77 78 79 80
## 3708.335 1877.746 2995.758 0.000 2211.200 0.000 0.000 0.000
## 81 82 83 84 85 86 87 88
## 2549.262 2122.862 2155.607 0.000 2337.344 3097.596 0.000 0.000
## 89 90 91 92 93 94 95 96
## 0.000 3477.726 0.000 0.000 0.000 0.000 0.000 0.000
## 97 98 99 100 101 102 103 104
## 0.000 0.000 2009.151 0.000 1958.144 2847.853 2695.691 3108.060
## 105 106 107 108 109 110 111 112
## 0.000 0.000 0.000 0.000 2626.702 0.000 2971.429 0.000
## 113 114 115 116 117 118 119 120
## 0.000 0.000 3267.391 0.000 0.000 3176.143 2584.099 0.000
## 121 122 123 124 125 126 127 128
## 0.000 3007.596 3359.524 2439.260 0.000 2328.712 2060.008 0.000
## 129 130 131 132 133 134 135 136
## 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
## 137 138 139 140 141 142 143 144
## 3742.517 4184.236 0.000 0.000 0.000 3615.936 0.000 0.000
## 145 146 147 148 149 150 151 152
## 0.000 2966.727 0.000 0.000 0.000 0.000 2885.267 2353.918
## 153 154 155 156 157 158 159 160
## 4034.475 2214.651 0.000 2579.792 0.000 0.000 2588.501 2626.691
## 161 162 163 164 165 166 167 168
## 0.000 0.000 0.000 0.000 3857.852 0.000 0.000 0.000
## 169 170 171 172 173 174 175 176
## 2100.014 0.000 0.000 3680.418 0.000 4635.774 2584.969 1807.019
## 177 178 179 180 181 182 183 184
## 3454.875 2702.984 3770.963 3269.976 3240.053 0.000 0.000 0.000
## 185 186 187 188 189 190 191 192
## 0.000 0.000 0.000 0.000 0.000 0.000 2240.906 3151.826
## 193 194 195 196 197 198 199 200
## 3512.474 0.000 0.000 2462.682 2566.310 0.000 0.000 0.000
## 201 202 203 204 205 206 207 208
## 0.000 2354.074 0.000 0.000 0.000 0.000 4483.710 0.000
## 209 210 211 212 213 214 215 216
## 0.000 0.000 0.000 0.000 2704.046 2882.920 0.000 0.000
## 217 218 219 220 221 222 223 224
## 2749.037 0.000 0.000 0.000 0.000 0.000 2517.245 1973.393
## 225 226 227 228 229 230 231 232
## 0.000 2025.719 3860.581 1696.430 2327.999 0.000 0.000 0.000
## 233 234 235 236 237 238 239 240
## 3922.886 0.000 0.000 0.000 0.000 0.000 0.000 3713.975
## 241 242 243 244 245 246 247 248
## 0.000 0.000 3961.794 0.000 0.000 2450.278 1669.442 2000.249
## 249 250 251 252 253 254 255 256
## 0.000 2858.901 2137.284 2595.073 0.000 2057.791 1884.467 1767.605
## 257 258 259 260 261 262 263 264
## 0.000 0.000 2321.565 0.000 0.000 0.000 0.000 0.000
## 265 266 267 268 269 270 271 272
## 0.000 0.000 0.000 0.000 4403.354 2268.714 2891.869 0.000
## 273 274 275 276 277 278 279 280
## 0.000 3849.188 0.000 0.000 4172.609 0.000 0.000 0.000
## 281 282 283 284 285 286 287 288
## 0.000 0.000 0.000 2184.814 2279.185 1776.982 0.000 1704.613
## 289 290 291 292 293 294 295 296
## 1686.425 3470.123 2190.371 0.000 0.000 2269.038 0.000 2213.298
## 297 298 299 300 301 302 303 304
## 0.000 2578.011 0.000 0.000 0.000 0.000 0.000 3338.808
## 305 306 307 308 309 310 311 312
## 2832.212 2245.462 0.000 2433.617 0.000 0.000 3750.491 0.000
## 313 314 315 316 317 318 319 320
## 0.000 4555.508 0.000 0.000 0.000 0.000 2431.446 0.000
## 321 322 323 324 325 326 327 328
## 0.000 2178.714 0.000 0.000 3153.108 2399.606 3058.715 0.000
## 329 330 331 332 333 334 335 336
## 3001.420 0.000 0.000 0.000 4064.832 0.000 0.000 0.000
## 337 338 339 340 341 342 343 344
## 0.000 1978.927 0.000 0.000 2369.255 3251.420 2291.295 3201.887
## 345 346 347 348 349 350 351 352
## 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
## 353 354 355 356 357 358 359 360
## 4127.216 3955.022 0.000 2783.045 2219.227 0.000 0.000 0.000
## 361 362 363 364 365 366 367 368
## 3653.111 0.000 0.000 1925.648 0.000 2955.017 0.000 1687.425
## 369 370 371 372 373 374 375 376
## 0.000 0.000 0.000 0.000 3005.873 0.000 0.000 3020.471
## 377 378 379 380 381 382 383 384
## 0.000 0.000 0.000 0.000 0.000 2453.288 0.000 0.000
## 385 386 387 388 389 390 391 392
## 0.000 2375.317 1951.675 0.000 0.000 2909.857 0.000 0.000
## 393 394 395 396 397 398 399 400
## 0.000 0.000 0.000 2626.056 0.000 1761.357 0.000 2671.869
## 401 402 403 404 405 406 407 408
## 2154.406 0.000 0.000 0.000 0.000 0.000 0.000 1948.084
## 409 410 411 412 413 414 415 416
## 0.000 0.000 0.000 3481.116 0.000 0.000 2477.336 1344.077
## 417 418 419 420 421 422 423 424
## 0.000 2038.607 0.000 0.000 2962.635 3229.071 3801.141 0.000
## 425 426 427 428 429 430 431 432
## 0.000 2034.676 0.000 0.000 2694.467 2200.606 1887.413 0.000
## 433 434 435 436 437 438 439 440
## 0.000 0.000 1797.262 3828.124 0.000 0.000 0.000 0.000
## 441 442 443 444 445 446 447 448
## 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
## 449 450 451 452 453 454 455 456
## 3386.994 4307.971 0.000 0.000 2401.326 0.000 2888.358 3131.317
## 457 458 459 460 461 462 463 464
## 3306.987 4173.593 0.000 1593.253 0.000 0.000 0.000 1669.183
## 465 466 467 468 469 470 471 472
## 0.000 0.000 3679.639 2492.097 0.000 0.000 0.000 3246.031
## 473 474 475 476 477 478 479 480
## 0.000 0.000 0.000 0.000 3746.304 4425.351 0.000 1979.157
## 481 482 483 484 485 486 487 488
## 0.000 0.000 0.000 2104.429 3634.381 2578.305 0.000 2678.231
## 489 490 491 492 493 494 495 496
## 0.000 3979.610 3342.514 0.000 0.000 0.000 0.000 2859.645
## 497 498 499 500 501 502 503 504
## 0.000 0.000 2318.816 0.000 0.000 0.000 4317.003 0.000
## 505 506 507 508 509 510 511 512
## 3461.419 0.000 2486.244 0.000 0.000 0.000 0.000 0.000
## 513 514 515 516 517 518 519 520
## 0.000 0.000 0.000 0.000 3425.586 3518.467 0.000 2381.040
## 521 522 523 524 525 526 527 528
## 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
## 529 530 531 532 533 534 535 536
## 0.000 0.000 0.000 0.000 0.000 1864.355 0.000 0.000
## 537 538 539 540 541 542 543 544
## 0.000 0.000 0.000 0.000 0.000 0.000 1911.985 3039.087
## 545 546 547 548 549 550 551 552
## 0.000 0.000 0.000 4084.071 2679.186 0.000 0.000 0.000
## 553 554 555 556 557 558 559 560
## 0.000 2005.854 0.000 0.000 1926.402 0.000 2352.524 0.000
## 561 562 563 564 565 566 567 568
## 0.000 1475.241 0.000 0.000 2341.011 0.000 5553.164 1742.953
## 569 570 571 572 573 574 575 576
## 2509.472 3443.720 0.000 0.000 0.000 0.000 0.000 0.000
## 577 578 579 580 581 582 583 584
## 0.000 0.000 2332.629 0.000 0.000 3118.198 0.000 3411.348
## 585 586 587 588 589 590 591 592
## 0.000 0.000 0.000 0.000 4418.533 2736.284 2304.623 0.000
## 593 594 595 596 597 598 599 600
## 0.000 2168.423 2452.577 2470.849 2682.927 0.000 0.000 2563.602
## 601 602 603 604 605 606 607 608
## 3187.197 0.000 0.000 0.000 2916.515 2462.188 2929.688 0.000
## 609 610 611 612 613 614 615 616
## 0.000 0.000 0.000 2624.340 0.000 0.000 2428.369 0.000
## 617 618 619 620 621 622 623 624
## 0.000 1913.871 0.000 3447.200 0.000 0.000 0.000 0.000
## 625 626 627 628 629 630 631 632
## 0.000 4331.933 2659.371 0.000 0.000 2950.859 0.000 0.000
## 633 634 635 636 637 638 639 640
## 0.000 0.000 0.000 0.000 0.000 3011.399 0.000 0.000
## 641 642 643 644 645 646 647 648
## 0.000 2555.876 0.000 0.000 0.000 3151.117 2463.916 0.000
## 649 650 651 652 653 654 655 656
## 0.000 0.000 0.000 0.000 3736.110 0.000 0.000 0.000
## 657 658 659 660 661 662 663 664
## 2837.473 0.000 0.000 0.000 2719.481 1480.734 1883.652 0.000
## 665 666 667 668 669 670 671 672
## 0.000 1706.428 2336.944 0.000 0.000 0.000 0.000 2810.806
## 673 674 675 676 677 678 679 680
## 3293.956 0.000 1655.906 0.000 0.000 2382.025 0.000 0.000
## 681 682 683 684 685 686 687 688
## 2872.617 0.000 2495.817 0.000 0.000 0.000 0.000 0.000
## 689 690 691 692 693 694 695 696
## 0.000 0.000 0.000 0.000 0.000 0.000 0.000 2482.088
## 697 698 699 700 701 702 703 704
## 0.000 2815.293 0.000 0.000 0.000 0.000 1950.378 0.000
## 705 706 707 708 709 710 711 712
## 0.000 0.000 1529.246 3939.748 0.000 0.000 0.000 0.000
## 713 714 715 716 717 718 719 720
## 2592.584 0.000 0.000 0.000 0.000 0.000 0.000 0.000
## 721 722 723 724 725 726 727 728
## 2725.121 0.000 0.000 0.000 0.000 0.000 0.000 0.000
## 729 730 731 732 733 734 735 736
## 0.000 0.000 2571.616 3372.828 0.000 0.000 2639.739 2722.785
## 737 738 739 740 741 742 743 744
## 0.000 0.000 0.000 0.000 3254.143 1727.746 3167.034 0.000
## 745 746 747 748 749 750 751 752
## 0.000 2435.198 3437.952 0.000 0.000 0.000 0.000 0.000
## 753 754 755 756 757 758 759 760
## 2679.807 2458.147 0.000 2082.780 2225.962 0.000 0.000 1586.105
## 761 762 763 764 765 766 767 768
## 0.000 4437.381 2712.816 3549.586 1796.978 2928.451 1517.131 0.000
## 769 770 771 772 773 774 775 776
## 2410.686 0.000 2472.165 2662.098 0.000 3451.139 0.000 2151.832
## 777 778 779 780 781 782 783 784
## 0.000 2240.260 0.000 0.000 0.000 3863.035 0.000 0.000
## 785 786 787 788 789 790 791 792
## 0.000 1907.128 0.000 0.000 0.000 0.000 0.000 0.000
## 793 794 795 796 797 798 799 800
## 3237.765 0.000 0.000 1712.911 0.000 2961.946 3414.876 2953.026
## 801 802 803 804 805 806 807 808
## 0.000 0.000 0.000 2172.330 0.000 0.000 0.000 0.000
## 809 810 811 812 813 814 815 816
## 0.000 2616.479 1936.291 0.000 0.000 0.000 0.000 0.000
## 817 818 819 820 821 822 823 824
## 0.000 3332.712 3057.742 0.000 2780.160 0.000 3421.097 2064.041
## 825 826 827 828 829 830 831 832
## 2551.820 0.000 0.000 0.000 0.000 0.000 0.000 0.000
## 833 834 835 836 837 838 839 840
## 3024.359 0.000 0.000 0.000 0.000 0.000 0.000 0.000
## 841 842 843 844 845 846 847 848
## 0.000 0.000 0.000 0.000 0.000 0.000 2203.552 2789.378
## 849 850 851 852 853 854 855 856
## 3413.111 3340.744 3318.875 0.000 0.000 0.000 2337.217 0.000
## 857 858 859 860 861 862 863 864
## 0.000 0.000 4089.994 0.000 0.000 3634.272 0.000 0.000
## 865 866 867 868 869 870 871 872
## 0.000 0.000 3214.805 0.000 0.000 2654.615 0.000 2861.403
## 873 874 875 876 877 878 879 880
## 0.000 4441.638 1732.851 0.000 0.000 0.000 0.000 2219.010
## 881 882 883 884 885 886 887 888
## 2422.483 0.000 0.000 0.000 3357.277 2330.594 4098.746 0.000
## 889 890 891 892 893 894 895 896
## 0.000 0.000 0.000 0.000 0.000 0.000 0.000 2271.413
## 897 898 899 900 901 902 903 904
## 0.000 0.000 1817.723 2768.293 0.000 0.000 4364.697 0.000
## 905 906 907 908 909 910 911 912
## 0.000 0.000 3346.777 0.000 2187.754 1896.713 3862.605 0.000
## 913 914 915 916 917 918 919 920
## 0.000 0.000 0.000 2220.237 5491.319 2680.660 0.000 1885.492
## 921 922 923 924 925 926 927 928
## 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
## 929 930 931 932 933 934 935 936
## 3682.021 0.000 0.000 4154.326 0.000 3235.422 0.000 2265.437
## 937 938 939 940 941 942 943 944
## 0.000 2083.855 3003.078 2057.612 3958.402 0.000 2106.087 0.000
## 945 946 947 948 949 950 951 952
## 3107.897 0.000 0.000 0.000 0.000 2413.498 0.000 0.000
## 953 954 955 956 957 958 959 960
## 0.000 2476.987 0.000 0.000 0.000 0.000 2814.717 0.000
## 961 962 963 964 965 966 967 968
## 2424.639 2756.560 2194.527 0.000 0.000 2942.710 2257.207 0.000
## 969 970 971 972 973 974 975 976
## 0.000 0.000 3043.575 0.000 0.000 2097.125 0.000 1439.480
## 977 978 979 980 981 982 983 984
## 1312.109 0.000 0.000 2587.057 0.000 0.000 3089.041 3221.927
## 985 986 987 988 989 990 991 992
## 3098.470 2673.620 2227.165 0.000 0.000 2398.908 0.000 0.000
## 993 994 995 996 997 998 999 1000
## 2534.730 0.000 2112.878 0.000 0.000 0.000 0.000 0.000
## 1001 1002 1003 1004 1005 1006 1007 1008
## 1533.552 3771.001 2933.645 0.000 2521.643 0.000 0.000 1771.994
## 1009 1010 1011 1012 1013 1014 1015 1016
## 2294.244 0.000 2912.976 0.000 0.000 0.000 0.000 2220.786
## 1017 1018 1019 1020 1021 1022 1023 1024
## 0.000 2304.818 1526.933 0.000 0.000 0.000 3640.500 2943.243
## 1025 1026 1027 1028 1029 1030 1031 1032
## 3611.271 3402.559 2102.239 0.000 0.000 0.000 0.000 0.000
## 1033 1034 1035 1036 1037 1038 1039 1040
## 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
## 1041 1042 1043 1044 1045 1046 1047 1048
## 0.000 2877.279 2246.896 0.000 3789.237 0.000 2776.471 2743.450
## 1049 1050 1051 1052 1053 1054 1055 1056
## 2607.127 2937.717 2606.392 2566.636 3330.434 0.000 2187.627 0.000
## 1057 1058 1059 1060 1061 1062 1063 1064
## 0.000 0.000 2434.113 3571.445 0.000 3276.093 2232.106 0.000
## 1065 1066 1067 1068 1069 1070 1071 1072
## 0.000 0.000 0.000 0.000 0.000 2293.489 0.000 0.000
## 1073 1074 1075 1076 1077 1078 1079 1080
## 0.000 4045.601 0.000 0.000 0.000 0.000 2717.972 0.000
## 1081 1082 1083 1084 1085 1086 1087 1088
## 3453.470 2942.332 0.000 0.000 2028.170 4824.958 0.000 0.000
## 1089 1090 1091 1092 1093 1094 1095 1096
## 0.000 0.000 0.000 0.000 0.000 2343.001 0.000 0.000
## 1097 1098 1099 1100 1101 1102 1103 1104
## 0.000 0.000 2225.471 3833.875 0.000 0.000 1735.579 2641.414
## 1105 1106 1107 1108 1109 1110 1111 1112
## 3368.086 1930.074 0.000 0.000 0.000 1605.138 2819.283 2778.051
## 1113 1114 1115 1116 1117 1118 1119 1120
## 0.000 0.000 0.000 0.000 2214.715 4301.324 2855.065 0.000
## 1121 1122 1123 1124 1125 1126 1127 1128
## 1925.772 2341.441 0.000 0.000 0.000 0.000 4740.345 0.000
## 1129 1130 1131 1132 1133 1134 1135 1136
## 0.000 0.000 0.000 0.000 3247.590 0.000 2559.589 0.000
## 1137 1138 1139 1140 1141 1142 1143 1144
## 0.000 0.000 0.000 0.000 0.000 0.000 0.000 3070.182
## 1145 1146 1147 1148 1149 1150 1151 1152
## 2468.054 0.000 0.000 3155.964 0.000 2905.352 2435.813 4610.611
## 1153 1154 1155 1156 1157 1158 1159 1160
## 0.000 1880.235 3248.764 0.000 0.000 0.000 0.000 0.000
## 1161 1162 1163 1164 1165 1166 1167 1168
## 0.000 0.000 0.000 0.000 0.000 2492.725 0.000 2442.304
## 1169 1170 1171 1172 1173 1174 1175 1176
## 1705.605 0.000 2218.092 2178.431 2548.757 2997.880 0.000 0.000
## 1177 1178 1179 1180 1181 1182 1183 1184
## 2440.670 0.000 0.000 3737.765 0.000 3895.382 0.000 2975.771
## 1185 1186 1187 1188 1189 1190 1191 1192
## 4344.651 0.000 0.000 0.000 0.000 0.000 1843.854 0.000
## 1193 1194 1195 1196 1197 1198 1199 1200
## 0.000 2821.233 0.000 0.000 0.000 0.000 2438.538 0.000
## 1201 1202 1203 1204 1205 1206 1207 1208
## 0.000 0.000 0.000 0.000 0.000 0.000 2382.325 0.000
## 1209 1210 1211 1212 1213 1214 1215 1216
## 0.000 2370.666 0.000 0.000 4456.245 0.000 0.000 0.000
## 1217 1218 1219 1220 1221 1222 1223 1224
## 2663.438 1832.310 0.000 0.000 0.000 0.000 3737.840 0.000
## 1225 1226 1227 1228 1229 1230 1231 1232
## 2731.563 0.000 0.000 0.000 0.000 4017.232 0.000 0.000
## 1233 1234 1235 1236 1237 1238 1239 1240
## 2424.866 2942.444 0.000 0.000 0.000 3178.158 0.000 0.000
## 1241 1242 1243 1244 1245 1246 1247 1248
## 1975.015 0.000 0.000 0.000 0.000 4427.320 3020.475 0.000
## 1249 1250 1251 1252 1253 1254 1255 1256
## 0.000 0.000 0.000 4208.323 0.000 2754.918 1832.689 0.000
## 1257 1258 1259 1260 1261 1262 1263 1264
## 3814.222 0.000 2983.326 0.000 0.000 0.000 0.000 4000.419
## 1265 1266 1267 1268 1269 1270 1271 1272
## 0.000 0.000 2509.787 0.000 2623.428 0.000 0.000 0.000
## 1273 1274 1275 1276 1277 1278 1279 1280
## 0.000 0.000 0.000 0.000 2796.126 0.000 0.000 0.000
## 1281 1282 1283 1284 1285 1286 1287 1288
## 2905.072 0.000 0.000 2156.525 3088.520 0.000 2912.690 0.000
## 1289 1290 1291 1292 1293 1294 1295 1296
## 0.000 0.000 2966.482 0.000 0.000 2366.823 2465.136 0.000
## 1297 1298 1299 1300 1301 1302 1303 1304
## 1837.770 0.000 2841.444 2112.169 0.000 0.000 0.000 0.000
## 1305 1306 1307 1308 1309 1310 1311 1312
## 0.000 0.000 3505.581 2744.253 0.000 3605.963 2164.576 0.000
## 1313 1314 1315 1316 1317 1318 1319 1320
## 4534.293 0.000 0.000 0.000 0.000 0.000 0.000 3912.146
## 1321 1322 1323 1324 1325 1326 1327 1328
## 0.000 2701.502 3612.022 0.000 0.000 2276.194 0.000 2129.110
## 1329 1330 1331 1332 1333 1334 1335 1336
## 0.000 0.000 0.000 0.000 0.000 0.000 1846.566 1826.129
## 1337 1338 1339 1340 1341 1342 1343 1344
## 0.000 0.000 2218.836 0.000 0.000 2541.135 0.000 0.000
## 1345 1346 1347 1348 1349 1350 1351 1352
## 4220.625 0.000 0.000 2404.862 0.000 0.000 0.000 2462.144
## 1353 1354 1355 1356 1357 1358 1359 1360
## 0.000 0.000 0.000 0.000 0.000 2658.243 0.000 0.000
## 1361 1362 1363 1364 1365 1366 1367 1368
## 0.000 0.000 4416.293 0.000 0.000 0.000 3019.589 2904.898
## 1369 1370 1371 1372 1373 1374 1375 1376
## 2505.151 3410.230 0.000 0.000 0.000 0.000 0.000 2387.049
## 1377 1378 1379 1380 1381 1382 1383 1384
## 2073.179 0.000 0.000 0.000 3397.011 3249.016 2520.349 0.000
## 1385 1386 1387 1388 1389 1390 1391 1392
## 0.000 0.000 0.000 0.000 0.000 2104.147 3007.173 2997.214
## 1393 1394 1395 1396 1397 1398 1399 1400
## 0.000 2308.643 0.000 0.000 0.000 2343.269 3082.486 0.000
## 1401 1402 1403 1404 1405 1406 1407 1408
## 0.000 0.000 3430.110 2301.070 1959.296 0.000 0.000 0.000
## 1409 1410 1411 1412 1413 1414 1415 1416
## 0.000 3167.317 0.000 0.000 0.000 0.000 0.000 0.000
## 1417 1418 1419 1420 1421 1422 1423 1424
## 0.000 0.000 2878.545 0.000 0.000 4503.386 2571.440 0.000
## 1425 1426 1427 1428 1429 1430 1431 1432
## 2870.248 4724.533 0.000 0.000 2480.895 0.000 0.000 0.000
## 1433 1434 1435 1436 1437 1438 1439 1440
## 0.000 0.000 0.000 2451.605 2180.356 2562.334 0.000 0.000
## 1441 1442 1443 1444 1445 1446 1447 1448
## 0.000 3594.218 2433.077 0.000 0.000 3233.921 3514.080 0.000
## 1449 1450 1451 1452 1453 1454 1455 1456
## 0.000 0.000 0.000 0.000 2730.898 0.000 0.000 0.000
## 1457 1458 1459 1460 1461 1462 1463 1464
## 3691.225 0.000 0.000 0.000 0.000 0.000 0.000 1941.113
## 1465 1466 1467 1468 1469 1470 1471 1472
## 0.000 2014.443 0.000 0.000 0.000 2273.432 0.000 0.000
## 1473 1474 1475 1476 1477 1478 1479 1480
## 0.000 1176.680 2281.769 0.000 2381.176 0.000 0.000 0.000
## 1481 1482 1483 1484 1485 1486 1487 1488
## 1735.304 0.000 1461.190 0.000 2480.721 0.000 0.000 0.000
## 1489 1490 1491 1492 1493 1494 1495 1496
## 3358.516 0.000 0.000 3673.179 0.000 0.000 3318.659 3308.538
## 1497 1498 1499 1500 1501 1502 1503 1504
## 0.000 0.000 0.000 0.000 0.000 2196.702 0.000 4101.940
## 1505 1506 1507 1508 1509 1510 1511 1512
## 0.000 0.000 2789.614 0.000 0.000 0.000 0.000 0.000
## 1513 1514 1515 1516 1517 1518 1519 1520
## 0.000 0.000 2566.604 3011.189 0.000 1816.189 0.000 0.000
## 1521 1522 1523 1524 1525 1526 1527 1528
## 2067.851 0.000 0.000 0.000 0.000 1660.241 0.000 2377.022
## 1529 1530 1531 1532 1533 1534 1535 1536
## 2559.583 3102.723 2716.465 2308.547 0.000 0.000 2096.965 0.000
## 1537 1538 1539 1540 1541 1542 1543 1544
## 0.000 2460.842 4159.926 3456.353 0.000 0.000 0.000 2152.571
## 1545 1546 1547 1548 1549 1550 1551 1552
## 3458.302 0.000 0.000 0.000 0.000 0.000 0.000 0.000
## 1553 1554 1555 1556 1557 1558 1559 1560
## 0.000 3076.298 0.000 0.000 0.000 0.000 2323.652 1707.102
## 1561 1562 1563 1564 1565 1566 1567 1568
## 0.000 3261.218 0.000 2723.979 4498.600 0.000 0.000 0.000
## 1569 1570 1571 1572 1573 1574 1575 1576
## 0.000 2124.674 3113.637 0.000 0.000 2528.814 0.000 2735.019
## 1577 1578 1579 1580 1581 1582 1583 1584
## 3062.032 0.000 0.000 0.000 0.000 0.000 0.000 0.000
## 1585 1586 1587 1588 1589 1590 1591 1592
## 0.000 0.000 0.000 1895.984 0.000 2434.607 1704.596 4131.673
## 1593 1594 1595 1596 1597 1598 1599 1600
## 0.000 0.000 0.000 0.000 0.000 0.000 0.000 2066.109
## 1601 1602 1603 1604 1605 1606 1607 1608
## 2623.772 0.000 1636.651 3093.027 0.000 4722.454 0.000 0.000
## 1609 1610 1611 1612 1613 1614 1615 1616
## 0.000 4460.783 1915.737 0.000 0.000 2714.232 2873.937 2348.252
## 1617 1618 1619 1620 1621 1622 1623 1624
## 0.000 2879.871 0.000 1603.527 2631.760 0.000 4117.763 0.000
## 1625 1626 1627 1628 1629 1630 1631 1632
## 0.000 0.000 0.000 0.000 0.000 4122.972 0.000 3136.958
## 1633 1634 1635 1636 1637 1638 1639 1640
## 0.000 1862.196 0.000 0.000 2435.631 0.000 0.000 0.000
## 1641 1642 1643 1644 1645 1646 1647 1648
## 0.000 0.000 0.000 0.000 2140.717 0.000 0.000 0.000
## 1649 1650 1651 1652 1653 1654 1655 1656
## 0.000 2829.234 0.000 0.000 0.000 0.000 3512.149 3173.811
## 1657 1658 1659 1660 1661 1662 1663 1664
## 0.000 0.000 0.000 0.000 1913.146 3267.732 3124.398 3460.127
## 1665 1666 1667 1668 1669 1670 1671 1672
## 5301.770 0.000 2250.503 0.000 4412.784 0.000 3024.591 0.000
## 1673 1674 1675 1676 1677 1678 1679 1680
## 3320.111 0.000 0.000 0.000 0.000 0.000 0.000 0.000
## 1681 1682 1683 1684 1685 1686 1687 1688
## 0.000 4064.803 3286.672 0.000 2892.279 0.000 0.000 2430.791
## 1689 1690 1691 1692 1693 1694 1695 1696
## 0.000 0.000 0.000 0.000 0.000 2913.360 1452.929 2989.230
## 1697 1698 1699 1700 1701 1702 1703 1704
## 0.000 3240.369 4221.373 0.000 2193.558 0.000 0.000 1693.800
## 1705 1706 1707 1708 1709 1710 1711 1712
## 0.000 0.000 3453.067 0.000 3408.903 0.000 0.000 0.000
## 1713 1714 1715 1716 1717 1718 1719 1720
## 2261.604 0.000 2574.320 0.000 0.000 0.000 0.000 0.000
## 1721 1722 1723 1724 1725 1726 1727 1728
## 0.000 1802.257 0.000 1970.695 1832.802 0.000 0.000 0.000
## 1729 1730 1731 1732 1733 1734 1735 1736
## 3208.159 2451.373 0.000 0.000 0.000 0.000 0.000 0.000
## 1737 1738 1739 1740 1741 1742 1743 1744
## 2021.554 0.000 0.000 0.000 4751.948 0.000 0.000 0.000
## 1745 1746 1747 1748 1749 1750 1751 1752
## 1753.775 0.000 0.000 0.000 2642.535 0.000 0.000 0.000
## 1753 1754 1755 1756 1757 1758 1759 1760
## 0.000 3111.269 0.000 0.000 0.000 2234.678 0.000 0.000
## 1761 1762 1763 1764 1765 1766 1767 1768
## 4068.571 0.000 0.000 0.000 2698.161 4051.058 0.000 0.000
## 1769 1770 1771 1772 1773 1774 1775 1776
## 0.000 0.000 0.000 0.000 0.000 4512.878 2208.011 0.000
## 1777 1778 1779 1780 1781 1782 1783 1784
## 1946.867 0.000 2841.342 0.000 0.000 0.000 0.000 1365.381
## 1785 1786 1787 1788 1789 1790 1791 1792
## 0.000 0.000 0.000 0.000 3903.320 2447.763 0.000 0.000
## 1793 1794 1795 1796 1797 1798 1799 1800
## 0.000 0.000 0.000 2546.477 0.000 0.000 0.000 2568.933
## 1801 1802 1803 1804 1805 1806 1807 1808
## 2369.726 0.000 0.000 0.000 0.000 2790.137 3115.469 3775.797
## 1809 1810 1811 1812 1813 1814 1815 1816
## 0.000 0.000 0.000 0.000 2971.247 0.000 2174.859 0.000
## 1817 1818 1819 1820 1821 1822 1823 1824
## 0.000 0.000 0.000 0.000 0.000 0.000 0.000 2015.685
## 1825 1826 1827 1828 1829 1830 1831 1832
## 0.000 3718.318 0.000 2554.984 0.000 0.000 0.000 0.000
## 1833 1834 1835 1836 1837 1838 1839 1840
## 0.000 0.000 2366.190 0.000 0.000 4302.100 0.000 3089.387
## 1841 1842 1843 1844 1845 1846 1847 1848
## 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
## 1849 1850 1851 1852 1853 1854 1855 1856
## 0.000 0.000 0.000 0.000 2884.313 0.000 0.000 0.000
## 1857 1858 1859 1860 1861 1862 1863 1864
## 0.000 0.000 0.000 0.000 2062.305 0.000 0.000 0.000
## 1865 1866 1867 1868 1869 1870 1871 1872
## 0.000 0.000 0.000 0.000 1764.117 2103.202 3259.153 0.000
## 1873 1874 1875 1876 1877 1878 1879 1880
## 0.000 0.000 2280.218 0.000 0.000 3919.511 2182.247 2384.420
## 1881 1882 1883 1884 1885 1886 1887 1888
## 0.000 0.000 0.000 0.000 1879.282 0.000 0.000 0.000
## 1889 1890 1891 1892 1893 1894 1895 1896
## 2751.730 0.000 0.000 0.000 0.000 0.000 2053.523 0.000
## 1897 1898 1899 1900 1901 1902 1903 1904
## 0.000 0.000 0.000 2481.630 0.000 3744.170 0.000 1934.822
## 1905 1906 1907 1908 1909 1910 1911 1912
## 0.000 0.000 0.000 0.000 2617.277 3718.700 0.000 2322.103
## 1913 1914 1915 1916 1917 1918 1919 1920
## 0.000 0.000 2666.930 0.000 0.000 0.000 0.000 0.000
## 1921 1922 1923 1924 1925 1926 1927 1928
## 0.000 2253.242 0.000 0.000 0.000 0.000 0.000 0.000
## 1929 1930 1931 1932 1933 1934 1935 1936
## 0.000 0.000 0.000 0.000 2879.233 3334.702 0.000 2532.755
## 1937 1938 1939 1940 1941 1942 1943 1944
## 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
## 1945 1946 1947 1948 1949 1950 1951 1952
## 0.000 0.000 3210.597 0.000 2782.624 0.000 0.000 0.000
## 1953 1954 1955 1956 1957 1958 1959 1960
## 2686.458 2230.628 0.000 0.000 0.000 0.000 0.000 2056.608
## 1961 1962 1963 1964 1965 1966 1967 1968
## 3550.316 2156.760 2187.033 0.000 0.000 0.000 0.000 3366.827
## 1969 1970 1971 1972 1973 1974 1975 1976
## 0.000 0.000 0.000 0.000 1726.972 3631.909 0.000 0.000
## 1977 1978 1979 1980 1981 1982 1983 1984
## 0.000 3560.464 1561.706 0.000 0.000 2463.568 0.000 0.000
## 1985 1986 1987 1988 1989 1990 1991 1992
## 0.000 0.000 0.000 0.000 0.000 2394.564 0.000 3274.981
## 1993 1994 1995 1996 1997 1998 1999 2000
## 3291.735 3006.880 0.000 0.000 2584.964 3871.384 0.000 0.000
## 2001 2002 2003 2004 2005 2006 2007 2008
## 2533.846 2760.668 3304.103 0.000 2895.439 0.000 3325.533 0.000
## 2009 2010 2011 2012 2013 2014 2015 2016
## 2217.715 2945.842 2104.206 0.000 5297.530 0.000 0.000 2596.098
## 2017 2018 2019 2020 2021 2022 2023 2024
## 0.000 3163.383 2940.468 0.000 0.000 0.000 1764.946 0.000
## 2025 2026 2027 2028 2029 2030 2031 2032
## 0.000 0.000 0.000 0.000 0.000 3365.061 0.000 0.000
## 2033 2034 2035 2036 2037 2038 2039 2040
## 0.000 0.000 2960.000 2393.535 0.000 0.000 2467.865 0.000
## 2041 2042 2043 2044 2045 2046 2047 2048
## 0.000 0.000 0.000 2333.033 0.000 0.000 0.000 2550.746
## 2049 2050 2051 2052 2053 2054 2055 2056
## 0.000 0.000 1997.586 0.000 3418.936 0.000 3103.695 0.000
## 2057 2058 2059 2060 2061 2062 2063 2064
## 0.000 0.000 0.000 0.000 0.000 1865.670 2390.485 0.000
## 2065 2066 2067 2068 2069 2070 2071 2072
## 0.000 0.000 3039.080 4384.116 0.000 0.000 0.000 0.000
## 2073 2074 2075 2076 2077 2078 2079 2080
## 4103.885 3175.578 0.000 0.000 4013.105 0.000 0.000 5420.134
## 2081 2082 2083 2084 2085 2086 2087 2088
## 0.000 0.000 2390.746 0.000 0.000 0.000 1894.218 2787.052
## 2089 2090 2091 2092 2093 2094 2095 2096
## 0.000 3019.695 0.000 0.000 2517.282 0.000 1884.067 3281.572
## 2097 2098 2099 2100 2101 2102 2103 2104
## 0.000 1591.509 3620.574 2632.414 2774.112 1736.851 3840.326 0.000
## 2105 2106 2107 2108 2109 2110 2111 2112
## 0.000 0.000 3282.669 0.000 0.000 0.000 4838.786 0.000
## 2113 2114 2115 2116 2117 2118 2119 2120
## 1423.687 0.000 0.000 0.000 3112.698 0.000 4873.935 0.000
## 2121 2122 2123 2124 2125 2126 2127 2128
## 0.000 0.000 3876.975 2718.690 0.000 0.000 2439.868 0.000
## 2129 2130 2131 2132 2133 2134 2135 2136
## 0.000 0.000 0.000 0.000 0.000 1683.902 3001.403 0.000
## 2137 2138 2139 2140 2141
## 0.000 0.000 0.000 0.000 0.000APPENDIX
library(ggplot2) insurance <- read.csv(“https://raw.githubusercontent.com/swigodsky/Data621/master/insurance_training_data.csv”, stringsAsFactors=FALSE) library(stringr)
head(insurance) nrow(insurance)
library(tidyr) library(dplyr)
insur_df <- insurance insur_df\(INCOME <- as.numeric(gsub("\\D+","",insur_df\)INCOME)) insur_df\(HOME_VAL <- as.numeric(gsub("\\D+","",insur_df\)HOME_VAL)) insur_df\(BLUEBOOK <- as.numeric(gsub("\\D+","",insur_df\)BLUEBOOK)) insur_df\(OLDCLAIM <- as.numeric(gsub("\\D+","",insur_df\)OLDCLAIM))
insur_df[insur_df==“No”]<-0 insur_df[insur_df==“Yes”]<-1 insur_df[insur_df==“no”]<-0 insur_df[insur_df==“yes”]<-1 insur_df\(MSTATUS[insur_df\)MSTATUS==“z_No”]<-0 insur_df\(SEX[insur_df\)SEX==“M”]<-0 insur_df\(SEX[insur_df\)SEX==“z_F”]<-1 insur_df\(CAR_USE[insur_df\)CAR_USE==“Private”]<-0 insur_df\(CAR_USE[insur_df\)CAR_USE==“Commercial”]<-1 insur_df\(URBANICITY[insur_df\)URBANICITY==“z_Highly Rural/ Rural”]<-0 insur_df\(URBANICITY[insur_df\)URBANICITY==“Highly Urban/ Urban”]<-1 insur_df\(PARENT1 <- as.numeric(insur_df\)PARENT1) insur_df\(MSTATUS <- as.numeric(insur_df\)MSTATUS) insur_df\(SEX <- as.numeric(insur_df\)SEX) insur_df\(CAR_USE <- as.numeric(insur_df\)CAR_USE) insur_df\(RED_CAR <- as.numeric(insur_df\)RED_CAR) insur_df\(REVOKED <- as.numeric(insur_df\)REVOKED) insur_df\(URBANICITY <- as.numeric(insur_df\)URBANICITY)
insur_df\(EDUCATION[insur_df\)EDUCATION == “
spec_val <- spec(test_df, test)
sens_val <- sens(test_df, test)
roc_tester <- rbind(roc_tester, list(o_m_specificity=1-spec_val,sensitivity= sens_val))#calculating Euclidean distance between point and y=x a2 <- c(1-spec_val,sens_val) b2 <- c(0,0) c2 <- c(1,1) d2 <- dist2d(a2,b2,c2) # distance of point a from line (b,c) if (d2>d){ d <- d2 cut_off_val <- cutoff }
#calculating area of trapezoid for each set of data points
if (cutoff>=0.1){
num_values = nrow(roc_tester)
base2 = roc_tester$sensitivity[num_values]
base1 = roc_tester$sensitivity[num_values-1]
height2 = roc_tester$o_m_specificity[num_values]
height1 = roc_tester$o_m_specificity[num_values-1]
area = .5*(base1+base2)*(height2-height1)
auc = auc + area
}}
roc_plot <- ggplot(roc_tester, aes(x = o_m_specificity, y = sensitivity)) + geom_point() + geom_abline(slope=1) + labs(x="False Positive Rate (1-specificity)", y="True Positive Rate (sensitivity)", title="ROC Curve" )
return(list(roc_plot=roc_plot, auc_val=auc, cut_off_val=cut_off_val)) }
pred_logit1 <- predict(logit1, newdata=test1, type=“response”)
roc_vals <- roc(pred_logit1,test1) roc_vals\(roc_plot print(roc_vals\)cut_off_val) print(roc_vals$auc_val)
pred_logit1[pred_logit1>=0.32] <- 1 pred_logit1[pred_logit1<0.32] <- 0
table(pred=round(pred_logit1), true=test1$TARGET_FLAG)
accuracy1 <- acc(round(pred_logit1), test1) error1 <- err(round(pred_logit1), test1) precision1 <- prec(round(pred_logit1), test1) sensitivity1 <- sens(round(pred_logit1), test1) specificity1 <- spec(round(pred_logit1), test1) f11 <- f1(round(pred_logit1), test1)
stat1 <- data.frame(list(accuracy=accuracy1, error=error1, precision=precision1, sensitivity=sensitivity1, specificity=specificity1, f1=f11)) print(stat1)
set.seed(15) n <- nrow(insur_df) shuffle_df2 <- insur_df[sample(n),] train_indeces <- 1:round(0.6n) train2 <- shuffle_df2[train_indeces,] test_indeces <- (round(.6n)+1):n test2 <- shuffle_df2[test_indeces,]
logit2 <- glm(train2$TARGET_FLAG ~ ., data=train2, family=binomial (link=“logit”)) logit2 <- update(logit2, .~. -TARGET_AMT, data = train2, family=binomial (link=“logit”)) logit2 <- update(logit2, .~. -INDEX, data = train2, family=binomial (link=“logit”)) summary(logit2)
logit2 <- update(logit2, .~. -z_Blue Collar, data = train2, family=binomial (link=“logit”)) summary(logit2)
logit2 <- update(logit2, .~. -Home Maker, data = train2, family=binomial (link=“logit”)) summary(logit2)
logit2 <- update(logit2, .~. -CAR_AGE, data = train2, family=binomial (link=“logit”)) summary(logit2)
logit2 <- update(logit2, .~. -SEX, data = train2, family=binomial (link=“logit”)) summary(logit2)
logit2 <- update(logit2, .~. -AGE, data = train2, family=binomial (link=“logit”)) summary(logit2)
logit2 <- update(logit2, .~. -Lawyer, data = train2, family=binomial (link=“logit”)) summary(logit2)
logit2 <- update(logit2, .~. -RED_CAR, data = train2, family=binomial (link=“logit”)) summary(logit2)
logit2 <- update(logit2, .~. -HOME_VAL, data = train2, family=binomial (link=“logit”)) summary(logit2)
logit2 <- update(logit2, .~. -Clerical, data = train2, family=binomial (link=“logit”)) summary(logit2)
logit2 <- update(logit2, .~. -HOMEKIDS, data = train2, family=binomial (link=“logit”)) summary(logit2)
logit2 <- update(logit2, .~. -YOJ, data = train2, family=binomial (link=“logit”)) summary(logit2)
logit2 <- update(logit2, .~. -OLDCLAIM, data = train2, family=binomial (link=“logit”)) summary(logit2)
logit2 <- update(logit2, .~. -Doctor, data = train2, family=binomial (link=“logit”)) summary(logit2) logitscalar2 <- mean(dlogis(predict(logit2,type=“link”))) logitscalar2*coef(logit2)
pred_logit2 <- predict(logit2, newdata=test2, type=“response”)
roc_vals2 <- roc(pred_logit2,test2) roc_vals2\(roc_plot print(roc_vals2\)cut_off_val) print(roc_vals2$auc_val)
pred_logit2[pred_logit2>=0.33] <- 1 pred_logit2[pred_logit2<0.33] <- 0
table(pred=round(pred_logit2), true=test2$TARGET_FLAG)
accuracy2 <- acc(round(pred_logit2), test2) error2 <- err(round(pred_logit2), test2) precision2 <- prec(round(pred_logit2), test2) sensitivity2 <- sens(round(pred_logit2), test2) specificity2 <- spec(round(pred_logit2), test2) f12 <- f1(round(pred_logit2), test2)
stat2 <- data.frame(list(accuracy=accuracy2, error=error2, precision=precision2, sensitivity=sensitivity2, specificity=specificity2, f1=f12)) print(stat2)
insur_df3 <-insur_df1
set.seed(1) n <- nrow(insur_df3) shuffle_df3 <- insur_df3[sample(n),] train_indeces <- 1:round(0.6n) train3 <- shuffle_df3[train_indeces,] test_indeces <- (round(.6n)+1):n test3 <- shuffle_df3[test_indeces,]
logit3 <- glm(train1$TARGET_FLAG ~ ., data=train3, family=binomial (link=“logit”)) logit3 <- update(logit3, .~. -TARGET_AMT, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -Panel Truck, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -oldclaim_freq, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -MSTATUS, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -homekids_age, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -BLUEBOOK, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -TIF, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -REVOKED, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -CAR_AGE, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -Van, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -Pickup, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -MVR_PTS, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -KIDSDRIV, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -YOJ, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -z_Blue Collar, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -sex_redcar_suv, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -Manager, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -Doctor, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -income_homeval_educ, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -Lawyer, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -TRAVTIME, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -CAR_USE, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -Professional, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -Sports Car, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -URBANICITY, data = train3, family=binomial (link=“logit”)) summary(logit3)
logit3 <- update(logit3, .~. -Home Maker, data = train3, family=binomial (link=“logit”)) logitscalar3 <- mean(dlogis(predict(logit3,type=“link”))) logitscalar3*coef(logit3)
pred_logit3 <- predict(logit3, newdata=test3, type=“response”)
roc_vals3 <- roc(pred_logit3,test3) roc_vals3\(roc_plot print(roc_vals3\)cut_off_val) print(roc_vals3$auc_val)
pred_logit3[pred_logit3>=0.23] <- 1 pred_logit3[pred_logit3<0.23] <- 0
table(pred=round(pred_logit3), true=test3$TARGET_FLAG)
accuracy3 <- acc(round(pred_logit3), test3) error3 <- err(round(pred_logit3), test3) precision3 <- prec(round(pred_logit3), test3) sensitivity3 <- sens(round(pred_logit3), test3) specificity3 <- spec(round(pred_logit3), test3) f13 <- f1(round(pred_logit3), test3)
stat3 <- data.frame(list(accuracy=accuracy3, error=error3, precision=precision3, sensitivity=sensitivity3, specificity=specificity3, f1=f13)) print(stat3)
train2a <- train2 train2a <- subset(train2a, select=-c(INDEX, TARGET_FLAG, z_Blue Collar, Home Maker, CAR_AGE, SEX, AGE, Lawyer, RED_CAR, HOME_VAL, Clerical, HOMEKIDS, YOJ, Doctor, OLDCLAIM))
claim_lm <- lm(train2a$TARGET_AMT ~., data=train2a) summary(claim_lm)
claim_lm <- update(claim_lm, .~. -Van, data = train2a) summary(claim_lm)
claim_lm <- update(claim_lm, .~. -BLUEBOOK, data = train2a) summary(claim_lm)
claim_lm <- update(claim_lm, .~. -Pickup, data = train2a) summary(claim_lm)
claim_lm <- update(claim_lm, .~. -CLM_FREQ, data = train2a) summary(claim_lm)
plot(fitted(claim_lm),resid(claim_lm)) qqnorm(resid(claim_lm)) qqline(resid(claim_lm))
predictclaim2a <- predict(claim_lm, newdata=test2, type=“response”) predictclaim2a[pred_logit2==0]<-0 predictclaim2a[predictclaim2a<0]<-0
error <- predictclaim2a-test2$TARGET_AMT
rmse <- sqrt(mean(error^2)) rmse
train2b <- train2 train2b <- subset(train2b, select=-c(INDEX, TARGET_FLAG, z_Blue Collar, Home Maker, CAR_AGE, SEX, AGE, Lawyer, RED_CAR, HOME_VAL, Clerical, HOMEKIDS, YOJ, Doctor, OLDCLAIM)) train2b[train2b==0] <- 0.000001 claim_lm2 <- lm(log(train2b$TARGET_AMT) ~., data=train2b) summary(claim_lm2)
plot(fitted(claim_lm2),resid(claim_lm2)) qqnorm(resid(claim_lm2)) qqline(resid(claim_lm2))
predictclaim2b <- predict(claim_lm2, newdata=test2, type=“response”) predictclaim2b <- exp(predictclaim2b) predictclaim2b[pred_logit2==0]<-0 predictclaim2b[predictclaim2b<0]<-0
error2 <- predictclaim2b-test2$TARGET_AMT
rmse2 <- sqrt(mean(error2^2)) rmse2
eval_data <- read.csv(‘https://raw.githubusercontent.com/swigodsky/Data621/master/insurance-evaluation-data.csv’, stringsAsFactors = FALSE)
eval_data\(INCOME <- as.numeric(gsub("\\D+","",eval_data\)INCOME)) eval_data\(HOME_VAL <- as.numeric(gsub("\\D+","",eval_data\)HOME_VAL)) eval_data\(BLUEBOOK <- as.numeric(gsub("\\D+","",eval_data\)BLUEBOOK)) eval_data\(OLDCLAIM <- as.numeric(gsub("\\D+","",eval_data\)OLDCLAIM))
eval_data[eval_data==“No”]<-0 eval_data[eval_data==“Yes”]<-1 eval_data[eval_data==“no”]<-0 eval_data[eval_data==“yes”]<-1 eval_data\(MSTATUS[eval_data\)MSTATUS==“z_No”]<-0 eval_data\(SEX[eval_data\)SEX==“M”]<-0 eval_data\(SEX[eval_data\)SEX==“z_F”]<-1 eval_data\(CAR_USE[eval_data\)CAR_USE==“Private”]<-0 eval_data\(CAR_USE[eval_data\)CAR_USE==“Commercial”]<-1 eval_data\(URBANICITY[eval_data\)URBANICITY==“z_Highly Rural/ Rural”]<-0 eval_data\(URBANICITY[eval_data\)URBANICITY==“Highly Urban/ Urban”]<-1 eval_data\(PARENT1 <- as.numeric(eval_data\)PARENT1) eval_data\(MSTATUS <- as.numeric(eval_data\)MSTATUS) eval_data\(SEX <- as.numeric(eval_data\)SEX) eval_data\(CAR_USE <- as.numeric(eval_data\)CAR_USE) eval_data\(RED_CAR <- as.numeric(eval_data\)RED_CAR) eval_data\(REVOKED <- as.numeric(eval_data\)REVOKED) eval_data\(URBANICITY <- as.numeric(eval_data\)URBANICITY)
eval_data\(EDUCATION[eval_data\)EDUCATION == “
predclm <- predict(claim_lm, newdata=eval_data, type=“response”) predclm[pred_flg==0]<-0 predclm[pred_flg<0]<-0 predclm