DATA 606 Final Project
Libraries
Introduction
Research question
About Company: Dream Housing Finance company deals in all home loans. They have presence across all urban, semi urban and rural areas. Customer first apply for home loan after that company validates the customer eligibility for loan.
Problem: Company wants to automate the loan eligibility process (real time) based on customer detail provided while filling online application form. These details are Gender, Marital Status, Education, Number of Dependents, Income, Loan Amount, Credit History and others. To automate this process, they have given a problem to identify the customers segments, those are eligible for loan amount so that they can specifically target these customers.
Data
This data source was given as part of a data science challenge or practice problem. I downloaded the data and loaded to my git-hub account. I will read the data into R from my git-hub account using raw link of the csv file using read.csv command.
Source: https://datahack.analyticsvidhya.com/contest/practice-problem-loan-prediction-iii/
# load data
my_loan_data<- read.csv("https://raw.githubusercontent.com/forhadakbar/data606fall2019stat/master/Final%20Project/Loan_prediction.csv")## Loan_ID Gender Married Dependents Education Self_Employed
## 1 LP001002 Male No 0 Graduate No
## 2 LP001003 Male Yes 1 Graduate No
## 3 LP001005 Male Yes 0 Graduate Yes
## 4 LP001006 Male Yes 0 Not Graduate No
## 5 LP001008 Male No 0 Graduate No
## 6 LP001011 Male Yes 2 Graduate Yes
## ApplicantIncome CoapplicantIncome LoanAmount Loan_Amount_Term
## 1 5849 0 NA 360
## 2 4583 1508 128 360
## 3 3000 0 66 360
## 4 2583 2358 120 360
## 5 6000 0 141 360
## 6 5417 4196 267 360
## Credit_History Property_Area Loan_Status
## 1 1 Urban Y
## 2 1 Rural N
## 3 1 Urban Y
## 4 1 Urban Y
## 5 1 Urban Y
## 6 1 Urban Y## [1] 614 13There are 614 cases and 13 columns. Each case or observation represent a loan application.
Exploratory Data Analysis & Inference
Dependent Variable
Loan_Status is the response variable. It is a categorical variable which gives us yes and no for loan approval status.
Independent Variable
I have few independent variables that i will consider for now. I will choose the most appropiate variables after doing exploratory analysis.
Applicants took a loan before. Credit history is the variable which answers that.
Applicants with higher incomes. So, we might look at the applicant income variable.
Applicants with higher education.
Gender of the applicant.
Number of Dependens an applicant has.
Property area contains location information of the loan property applied for.
Relevant summary statistics
## 'data.frame': 614 obs. of 13 variables:
## $ Loan_ID : Factor w/ 614 levels "LP001002","LP001003",..: 1 2 3 4 5 6 7 8 9 10 ...
## $ Gender : Factor w/ 3 levels "","Female","Male": 3 3 3 3 3 3 3 3 3 3 ...
## $ Married : Factor w/ 3 levels "","No","Yes": 2 3 3 3 2 3 3 3 3 3 ...
## $ Dependents : Factor w/ 5 levels "","0","1","2",..: 2 3 2 2 2 4 2 5 4 3 ...
## $ Education : Factor w/ 2 levels "Graduate","Not Graduate": 1 1 1 2 1 1 2 1 1 1 ...
## $ Self_Employed : Factor w/ 3 levels "","No","Yes": 2 2 3 2 2 3 2 2 2 2 ...
## $ ApplicantIncome : int 5849 4583 3000 2583 6000 5417 2333 3036 4006 12841 ...
## $ CoapplicantIncome: num 0 1508 0 2358 0 ...
## $ LoanAmount : int NA 128 66 120 141 267 95 158 168 349 ...
## $ Loan_Amount_Term : int 360 360 360 360 360 360 360 360 360 360 ...
## $ Credit_History : int 1 1 1 1 1 1 1 0 1 1 ...
## $ Property_Area : Factor w/ 3 levels "Rural","Semiurban",..: 3 1 3 3 3 3 3 2 3 2 ...
## $ Loan_Status : Factor w/ 2 levels "N","Y": 2 1 2 2 2 2 2 1 2 1 ...## Loan_ID Gender Married Dependents Education
## LP001002: 1 : 13 : 3 : 15 Graduate :480
## LP001003: 1 Female:112 No :213 0 :345 Not Graduate:134
## LP001005: 1 Male :489 Yes:398 1 :102
## LP001006: 1 2 :101
## LP001008: 1 3+: 51
## LP001011: 1
## (Other) :608
## Self_Employed ApplicantIncome CoapplicantIncome LoanAmount
## : 32 Min. : 150 Min. : 0 Min. : 9.0
## No :500 1st Qu.: 2878 1st Qu.: 0 1st Qu.:100.0
## Yes: 82 Median : 3812 Median : 1188 Median :128.0
## Mean : 5403 Mean : 1621 Mean :146.4
## 3rd Qu.: 5795 3rd Qu.: 2297 3rd Qu.:168.0
## Max. :81000 Max. :41667 Max. :700.0
## NA's :22
## Loan_Amount_Term Credit_History Property_Area Loan_Status
## Min. : 12 Min. :0.0000 Rural :179 N:192
## 1st Qu.:360 1st Qu.:1.0000 Semiurban:233 Y:422
## Median :360 Median :1.0000 Urban :202
## Mean :342 Mean :0.8422
## 3rd Qu.:360 3rd Qu.:1.0000
## Max. :480 Max. :1.0000
## NA's :14 NA's :50Data Cleaning
LoanAmount variable has 22 Null Value -Loan_Amount_Term has 14 null values -Credit_History has 50 Null values Data set observation.
#Store backup before removing missing values
my_loan_data_backup <- my_loan_data
#Retrun all rows with missing values
my_loan_data[!complete.cases(my_loan_data),]## Loan_ID Gender Married Dependents Education Self_Employed
## 1 LP001002 Male No 0 Graduate No
## 17 LP001034 Male No 1 Not Graduate No
## 20 LP001041 Male Yes 0 Graduate
## 25 LP001052 Male Yes 1 Graduate
## 31 LP001091 Male Yes 1 Graduate
## 36 LP001106 Male Yes 0 Graduate No
## 37 LP001109 Male Yes 0 Graduate No
## 43 LP001123 Male Yes 0 Graduate No
## 45 LP001136 Male Yes 0 Not Graduate Yes
## 46 LP001137 Female No 0 Graduate No
## 64 LP001213 Male Yes 1 Graduate No
## 74 LP001250 Male Yes 3+ Not Graduate No
## 80 LP001264 Male Yes 3+ Not Graduate Yes
## 82 LP001266 Male Yes 1 Graduate Yes
## 84 LP001273 Male Yes 0 Graduate No
## 87 LP001280 Male Yes 2 Not Graduate No
## 96 LP001326 Male No 0 Graduate
## 103 LP001350 Male Yes Graduate No
## 104 LP001356 Male Yes 0 Graduate No
## 113 LP001391 Male Yes 0 Not Graduate No
## 114 LP001392 Female No 1 Graduate Yes
## 118 LP001405 Male Yes 1 Graduate No
## 126 LP001443 Female No 0 Graduate No
## 128 LP001449 Male No 0 Graduate No
## 130 LP001465 Male Yes 0 Graduate No
## 131 LP001469 Male No 0 Graduate Yes
## 157 LP001541 Male Yes 1 Graduate No
## 166 LP001574 Male Yes 0 Graduate No
## 182 LP001634 Male No 0 Graduate No
## 188 LP001643 Male Yes 0 Graduate No
## 198 LP001669 Female No 0 Not Graduate No
## 199 LP001671 Female Yes 0 Graduate No
## 203 LP001682 Male Yes 3+ Not Graduate No
## 220 LP001734 Female Yes 2 Graduate No
## 224 LP001749 Male Yes 0 Graduate No
## 233 LP001770 Male No 0 Not Graduate No
## 237 LP001786 Male Yes 0 Graduate
## 238 LP001788 Female No 0 Graduate Yes
## 260 LP001864 Male Yes 3+ Not Graduate No
## 261 LP001865 Male Yes 1 Graduate No
## 280 LP001908 Female Yes 0 Not Graduate No
## 285 LP001922 Male Yes 0 Graduate No
## 306 LP001990 Male No 0 Not Graduate No
## 310 LP001998 Male Yes 2 Not Graduate No
## 314 LP002008 Male Yes 2 Graduate Yes
## 318 LP002036 Male Yes 0 Graduate No
## 319 LP002043 Female No 1 Graduate No
## 323 LP002054 Male Yes 2 Not Graduate No
## 324 LP002055 Female No 0 Graduate No
## 336 LP002106 Male Yes Graduate Yes
## 339 LP002113 Female No 3+ Not Graduate No
## 349 LP002137 Male Yes 0 Graduate No
## 364 LP002178 Male Yes 0 Graduate No
## 368 LP002188 Male No 0 Graduate No
## 378 LP002223 Male Yes 0 Graduate No
## 388 LP002243 Male Yes 0 Not Graduate No
## 393 LP002263 Male Yes 0 Graduate No
## 396 LP002272 Male Yes 2 Graduate No
## 412 LP002319 Male Yes 0 Graduate
## 422 LP002357 Female No 0 Not Graduate No
## 424 LP002362 Male Yes 1 Graduate No
## 436 LP002393 Female Graduate No
## 438 LP002401 Male Yes 0 Graduate No
## 445 LP002424 Male Yes 0 Graduate No
## 450 LP002444 Male No 1 Not Graduate Yes
## 452 LP002447 Male Yes 2 Not Graduate No
## 461 LP002478 Yes 0 Graduate Yes
## 474 LP002522 Female No 0 Graduate Yes
## 480 LP002533 Male Yes 2 Graduate No
## 491 LP002560 Male No 0 Not Graduate No
## 492 LP002562 Male Yes 1 Not Graduate No
## 498 LP002588 Male Yes 0 Graduate No
## 504 LP002618 Male Yes 1 Not Graduate No
## 507 LP002624 Male Yes 0 Graduate No
## 525 LP002697 Male No 0 Graduate No
## 531 LP002717 Male Yes 0 Graduate No
## 534 LP002729 Male No 1 Graduate No
## 545 LP002757 Female Yes 0 Not Graduate No
## 551 LP002778 Male Yes 2 Graduate Yes
## 552 LP002784 Male Yes 1 Not Graduate No
## 557 LP002794 Female No 0 Graduate No
## 566 LP002833 Male Yes 0 Not Graduate No
## 584 LP002898 Male Yes 1 Graduate No
## 601 LP002949 Female No 3+ Graduate
## 606 LP002960 Male Yes 0 Not Graduate No
## ApplicantIncome CoapplicantIncome LoanAmount Loan_Amount_Term
## 1 5849 0 NA 360
## 17 3596 0 100 240
## 20 2600 3500 115 NA
## 25 3717 2925 151 360
## 31 4166 3369 201 360
## 36 2275 2067 NA 360
## 37 1828 1330 100 NA
## 43 2400 0 75 360
## 45 4695 0 96 NA
## 46 3410 0 88 NA
## 64 4945 0 NA 360
## 74 4755 0 95 NA
## 80 3333 2166 130 360
## 82 2395 0 NA 360
## 84 6000 2250 265 360
## 87 3333 2000 99 360
## 96 6782 0 NA 360
## 103 13650 0 NA 360
## 104 4652 3583 NA 360
## 113 3572 4114 152 NA
## 114 7451 0 NA 360
## 118 2214 1398 85 360
## 126 3692 0 93 360
## 128 3865 1640 NA 360
## 130 6080 2569 182 360
## 131 20166 0 650 480
## 157 6000 0 160 360
## 166 3707 3166 182 NA
## 182 1916 5063 67 360
## 188 2383 2138 58 360
## 198 1907 2365 120 NA
## 199 3416 2816 113 360
## 203 3992 0 NA 180
## 220 4283 2383 127 360
## 224 7578 1010 175 NA
## 233 3189 2598 120 NA
## 237 5746 0 255 360
## 238 3463 0 122 360
## 260 4931 0 128 360
## 261 6083 4250 330 360
## 280 4100 0 124 360
## 285 20667 0 NA 360
## 306 2000 0 NA 360
## 310 7667 0 185 360
## 314 5746 0 144 84
## 318 2058 2134 88 360
## 319 3541 0 112 360
## 323 3601 1590 NA 360
## 324 3166 2985 132 360
## 336 5503 4490 70 NA
## 339 1830 0 NA 360
## 349 6333 4583 259 360
## 364 3013 3033 95 300
## 368 5124 0 124 NA
## 378 4310 0 130 360
## 388 3010 3136 NA 360
## 393 2583 2115 120 360
## 396 3276 484 135 360
## 412 6256 0 160 360
## 422 2720 0 80 NA
## 424 7250 1667 110 NA
## 436 10047 0 NA 240
## 438 2213 1125 NA 360
## 445 7333 8333 175 300
## 450 2769 1542 190 360
## 452 1958 1456 60 300
## 461 2083 4083 160 360
## 474 2500 0 93 360
## 480 2947 1603 NA 360
## 491 2699 2785 96 360
## 492 5333 1131 186 360
## 498 4625 2857 111 12
## 504 4050 5302 138 360
## 507 20833 6667 480 360
## 525 4680 2087 NA 360
## 531 1025 5500 216 360
## 534 11250 0 196 360
## 545 3017 663 102 360
## 551 6633 0 NA 360
## 552 2492 2375 NA 360
## 557 2667 1625 84 360
## 566 4467 0 120 360
## 584 1880 0 61 360
## 601 416 41667 350 180
## 606 2400 3800 NA 180
## Credit_History Property_Area Loan_Status
## 1 1 Urban Y
## 17 NA Urban Y
## 20 1 Urban Y
## 25 NA Semiurban N
## 31 NA Urban N
## 36 1 Urban Y
## 37 0 Urban N
## 43 NA Urban Y
## 45 1 Urban Y
## 46 1 Urban Y
## 64 0 Rural N
## 74 0 Semiurban N
## 80 NA Semiurban Y
## 82 1 Semiurban Y
## 84 NA Semiurban N
## 87 NA Semiurban Y
## 96 NA Urban N
## 103 1 Urban Y
## 104 1 Semiurban Y
## 113 0 Rural N
## 114 1 Semiurban Y
## 118 NA Urban Y
## 126 NA Rural Y
## 128 1 Rural Y
## 130 NA Rural N
## 131 NA Urban Y
## 157 NA Rural Y
## 166 1 Rural Y
## 182 NA Rural N
## 188 NA Rural Y
## 198 1 Urban Y
## 199 NA Semiurban Y
## 203 1 Urban N
## 220 NA Semiurban Y
## 224 1 Semiurban Y
## 233 1 Rural Y
## 237 NA Urban N
## 238 NA Urban Y
## 260 NA Semiurban N
## 261 NA Urban Y
## 280 NA Rural Y
## 285 1 Rural N
## 306 1 Urban N
## 310 NA Rural Y
## 314 NA Rural Y
## 318 NA Urban Y
## 319 NA Semiurban Y
## 323 1 Rural Y
## 324 NA Rural Y
## 336 1 Semiurban Y
## 339 0 Urban N
## 349 NA Semiurban Y
## 364 NA Urban Y
## 368 0 Rural N
## 378 NA Semiurban Y
## 388 0 Urban N
## 393 NA Urban Y
## 396 NA Semiurban Y
## 412 NA Urban Y
## 422 0 Urban N
## 424 0 Urban N
## 436 1 Semiurban Y
## 438 1 Urban Y
## 445 NA Rural Y
## 450 NA Semiurban N
## 452 NA Urban Y
## 461 NA Semiurban Y
## 474 NA Urban Y
## 480 1 Urban N
## 491 NA Semiurban Y
## 492 NA Urban Y
## 498 NA Urban Y
## 504 NA Rural N
## 507 NA Urban Y
## 525 1 Semiurban N
## 531 NA Rural Y
## 534 NA Semiurban N
## 545 NA Semiurban Y
## 551 0 Rural N
## 552 1 Rural Y
## 557 NA Urban Y
## 566 NA Rural Y
## 584 NA Rural N
## 601 NA Urban N
## 606 1 Urban NVisual Analysis
Property Area:
## Rural Semiurban Urban
## 155 209 165ggplot(data=my_loan_data, aes(my_loan_data$Property_Area)) +
geom_histogram(col="red",fill="lightblue",stat="count" ) +
facet_grid(~my_loan_data$Loan_Status)+
scale_x_discrete()## Warning: Ignoring unknown parameters: binwidth, bins, pad
Histogram of Property Area shows that Loan approval is more into Semiurban area than Rural and Urban. Urban area has lowest loan approval. Loan rejection is lowest in Rural area. Semiurban & Urban has same loan rejection
Coapplicant Income:
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0 0 1086 1542 2232 33837ggplot(data=my_loan_data, aes(x= my_loan_data$CoapplicantIncome)) +
geom_histogram(col="red",fill="lightblue", bins = 15) +
facet_grid(~my_loan_data$Loan_Status)+
theme_bw()
Histogram shows that low income peoples are mainly applying for loans and number of loan rejection is more in the lowest income segment
Education:
## Graduate Not Graduate
## 421 108ggplot(data=my_loan_data, aes(my_loan_data$Education)) +
geom_histogram(col="red",fill="lightblue",stat="count" ) +
facet_grid(~my_loan_data$Loan_Status)+
scale_x_discrete()+
theme_bw()## Warning: Ignoring unknown parameters: binwidth, bins, pad
Based on loan approval flag shows that - loan approval rate for graduate is more than non graduate
Number of Dependents:
## 0 1 2 3+
## 12 295 85 92 45ggplot(data=my_loan_data, aes(my_loan_data$Dependents)) +
geom_histogram(col="red",fill="lightblue",stat="count" ) +
facet_grid(~my_loan_data$Loan_Status)+
scale_x_discrete()+
theme_bw()## Warning: Ignoring unknown parameters: binwidth, bins, pad
Loan approval shows that -People having no dependents have maximum loan approval and rejection count
Gender:
## Female Male
## 12 95 422ggplot(data=my_loan_data, aes(my_loan_data$Gender)) +
geom_histogram(col="red",fill="lightblue",stat="count") +
facet_grid(~my_loan_data$Loan_Status)+
scale_x_discrete()+
theme_bw()## Warning: Ignoring unknown parameters: binwidth, bins, pad
Male applicant has higher loan approval and rejection count than female applicant. So this looks to be an influencing factor
Logestic Regression
Logistic Regression, in simple terms, predicts the probability of occurrence of an event by fitting data to a logit function. Regression coefficients represent the mean change in the response variable for one unit of change in the predictor variable while holding other predictors in the model constant. This type of models is part of a larger class of algorithms known as Generalized Linear Model or GLM.
Preparing Data for The Model:
Logistic Regression Model
##
## Call:
## glm(formula = Loan_Status ~ ., family = "binomial", data = traindf)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.4568 -0.3237 0.4808 0.6846 2.5320
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.081e+01 6.111e+02 0.018 0.98589
## GenderFemale 2.149e-01 8.331e-01 0.258 0.79642
## GenderMale 4.826e-01 7.696e-01 0.627 0.53058
## MarriedNo -1.411e+01 6.111e+02 -0.023 0.98158
## MarriedYes -1.357e+01 6.111e+02 -0.022 0.98228
## Dependents0 8.100e-01 1.192e+00 0.679 0.49685
## Dependents1 4.394e-01 1.203e+00 0.365 0.71489
## Dependents2 8.817e-01 1.210e+00 0.729 0.46605
## Dependents3+ 1.233e+00 1.272e+00 0.970 0.33226
## EducationNot Graduate -5.371e-01 3.382e-01 -1.588 0.11231
## Self_EmployedNo -5.074e-01 6.199e-01 -0.819 0.41301
## Self_EmployedYes -5.537e-01 6.876e-01 -0.805 0.42069
## ApplicantIncome -2.080e-06 2.857e-05 -0.073 0.94194
## CoapplicantIncome -2.636e-05 5.068e-05 -0.520 0.60296
## LoanAmount -6.392e-04 1.964e-03 -0.325 0.74483
## Loan_Amount_Term -1.527e-03 2.373e-03 -0.643 0.51993
## Credit_History 4.069e+00 5.172e-01 7.868 3.61e-15 ***
## Property_AreaSemiurban 1.118e+00 3.413e-01 3.275 0.00105 **
## Property_AreaUrban -1.863e-02 3.247e-01 -0.057 0.95426
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 521.94 on 422 degrees of freedom
## Residual deviance: 367.84 on 404 degrees of freedom
## AIC: 405.84
##
## Number of Fisher Scoring iterations: 13Most significant variables are
- Credit_History
- Property_AreaSemiurban
## 4 14 18 30 39 40
## 0.69760640 0.69986730 0.03188838 0.90789019 0.65999881 0.81105889
## 44 47 49 53 56 60
## 0.92402335 0.75153841 0.04776744 0.85421239 0.93258247 0.70629934
## 63 65 72 73 92 93
## 0.04549148 0.09082769 0.88357643 0.89554146 0.93808251 0.72017125
## 97 116 119 132 135 136
## 0.90314301 0.78440827 0.79510727 0.70625984 0.88500588 0.93577745
## 139 142 145 147 149 155
## 0.10536622 0.70286357 0.84810181 0.77654341 0.62762369 0.71928099
## 162 165 169 186 202 205
## 0.79001026 0.84873773 0.10145128 0.91085452 0.11944255 0.89099349
## 210 212 215 217 225 241
## 0.88368575 0.24711231 0.79305940 0.80247664 0.92817051 0.80011242
## 248 257 259 263 267 268
## 0.68634336 0.02426589 0.77828094 0.86601522 0.92977344 0.09214989
## 278 284 287 294 296 301
## 0.80618558 0.65243304 0.85567225 0.01150891 0.91177597 0.03806886
## 313 317 321 326 332 333
## 0.66587059 0.93345498 0.92540217 0.03422976 0.88567936 0.52956330
## 341 342 346 348 351 356
## 0.77973624 0.66735082 0.92283297 0.71254337 0.92337455 0.52365996
## 366 387 389 391 403 407
## 0.58981379 0.80962140 0.79604799 0.78316299 0.82022176 0.69822926
## 408 418 421 428 429 430
## 0.53660313 0.70253745 0.80776644 0.93264766 0.81363255 0.66224382
## 435 440 469 470 473 481
## 0.71257139 0.64147570 0.90788824 0.79791876 0.95272422 0.52148145
## 487 490 496 499 502 510
## 0.04488503 0.71638938 0.91203650 0.89753808 0.85544079 0.57156303
## 511 516 521 541 546 548
## 0.58802588 0.85392337 0.89178412 0.89247082 0.79893656 0.70126172
## 567 573 574 576 585 587
## 0.71665907 0.79875462 0.86176548 0.87242612 0.04506690 0.80311795
## 596 598 602 610
## 0.59456155 0.05596149 0.79042980 0.66355147## Predictedvalue
## Actualvalue FALSE TRUE
## N 14 19
## Y 2 71## [1] 0.8207547Accuracy: 82.07%
Decision Tree
Decision trees create a set of binary splits on the predictor variables in order to create a tree that can be used to classify new observations into one of two groups. Here, we will be using classical trees. The algorithm of this model is the following:
Choose the predictor variable that best splits the data into two groups;
Separate the data into these two groups;
Repeat these steps until a subgroup contains fewer than a minimum number of observations;
To classify a case, run it down the tree to a terminal node, and assign it the model outcome value assigned in the previous step.
set.seed(42)
sample <- sample.int(n = nrow(my_loan_data_1), size = floor(.70*nrow(my_loan_data_1)), replace = F)
trainnew <- my_loan_data_1[sample, ]
testnew <- my_loan_data_1[-sample, ]dtree <- rpart(Loan_Status ~ Credit_History + Education + Self_Employed + Property_Area + LoanAmount +
ApplicantIncome, method="class", data=traindf,parms=list(split="information"))
dtree$cptable## CP nsplit rel error xerror xstd
## 1 0.40769231 0 1.0000000 1.0000000 0.07299480
## 2 0.01346154 1 0.5923077 0.5923077 0.06104778
## 3 0.01000000 5 0.5384615 0.6384615 0.06282970
dtree.pruned <- prune(dtree, cp=.02290076)
library(rpart.plot)
prp(dtree.pruned, type = 2, extra = 104,
fallen.leaves = TRUE, main="Decision Tree")
dtree.pred <- predict(dtree.pruned, trainnew, type="class")
dtree.perf <- table(trainnew$Loan_Status, dtree.pred,
dnn=c("Actual", "Predicted"))
dtree.perf## Predicted
## Actual N Y
## N 49 64
## Y 5 252dtree_test <- rpart(Loan_Status ~ Credit_History+Education+Self_Employed+Property_Area+LoanAmount+
ApplicantIncome,method="class", data=testnew,parms=list(split="information"))
dtree_test$cptable## CP nsplit rel error xerror xstd
## 1 0.42 0 1.00 1.00 0.11709266
## 2 0.01 1 0.58 0.58 0.09738725
dtree_test.pruned <- prune(dtree_test, cp=.01639344)
prp(dtree_test.pruned, type = 2, extra = 104,
fallen.leaves = TRUE, main="Decision Tree")
Accuracy: 84% Results show better performance than the logistic model.
Random Forest
set.seed(42)
fit.forest <- randomForest(Loan_Status ~ Credit_History+Education+Self_Employed+Property_Area+LoanAmount+
ApplicantIncome, data=trainnew,
na.action=na.roughfix,
importance=TRUE)
fit.forest##
## Call:
## randomForest(formula = Loan_Status ~ Credit_History + Education + Self_Employed + Property_Area + LoanAmount + ApplicantIncome, data = trainnew, importance = TRUE, na.action = na.roughfix)
## Type of random forest: classification
## Number of trees: 500
## No. of variables tried at each split: 2
##
## OOB estimate of error rate: 18.65%
## Confusion matrix:
## N Y class.error
## N 53 60 0.53097345
## Y 9 248 0.03501946## MeanDecreaseGini
## Credit_History 41.890029
## Education 3.328873
## Self_Employed 4.921181
## Property_Area 8.718707
## LoanAmount 29.449887
## ApplicantIncome 29.320255forest.pred <- predict(fit.forest, testnew)
forest.perf <- table(testnew$Loan_Status, forest.pred,
dnn=c("Actual", "Predicted"))
forest.perf## Predicted
## Actual N Y
## N 24 26
## Y 5 104Here is the accuracy of the model: 80.50%
Conclusion
After analyzing the data from the loan prediction dataset, the data shows that Credit History and Property_AreaSemiurban are most significant variables to predict whether a loan application will approved or not. We can predict the loan approval using different models. Here, we got 82.07% accuracy for logistic regresission, 84% accuracy for Decesion tree and 80.50% accuracy for random forest.
The dataset is relatively small. A larger dataset will help to improve the model accuracy.
We can conclude that the company should target customers with Credit history and customer who lives in Semiurban area.