Mid-Term
Ochirpurev
2023-04-14
Question1 (A) To load and display the descriptive statistics of the mtcars data :
## vars n mean sd median trimmed mad min max range skew
## mpg 1 32 20.09 6.03 19.20 19.70 5.41 10.40 33.90 23.50 0.61
## cyl 2 32 6.19 1.79 6.00 6.23 2.97 4.00 8.00 4.00 -0.17
## disp 3 32 230.72 123.94 196.30 222.52 140.48 71.10 472.00 400.90 0.38
## hp 4 32 146.69 68.56 123.00 141.19 77.10 52.00 335.00 283.00 0.73
## drat 5 32 3.60 0.53 3.70 3.58 0.70 2.76 4.93 2.17 0.27
## wt 6 32 3.22 0.98 3.33 3.15 0.77 1.51 5.42 3.91 0.42
## qsec 7 32 17.85 1.79 17.71 17.83 1.42 14.50 22.90 8.40 0.37
## vs 8 32 0.44 0.50 0.00 0.42 0.00 0.00 1.00 1.00 0.24
## am 9 32 0.41 0.50 0.00 0.38 0.00 0.00 1.00 1.00 0.36
## gear 10 32 3.69 0.74 4.00 3.62 1.48 3.00 5.00 2.00 0.53
## carb 11 32 2.81 1.62 2.00 2.65 1.48 1.00 8.00 7.00 1.05
## kurtosis se
## mpg -0.37 1.07
## cyl -1.76 0.32
## disp -1.21 21.91
## hp -0.14 12.12
## drat -0.71 0.09
## wt -0.02 0.17
## qsec 0.34 0.32
## vs -2.00 0.09
## am -1.92 0.09
## gear -1.07 0.13
## carb 1.26 0.29## Min. 1st Qu. Median Mean 3rd Qu. Max. sds se
## mpg 10.400 15.42500 19.200 20.090625 22.80 33.900 6.0269481 4.2616958
## cyl 4.000 4.00000 6.000 6.187500 8.00 8.000 1.7859216 1.2628373
## disp 71.100 120.82500 196.300 230.721875 326.00 472.000 123.9386938 87.6378909
## hp 52.000 96.50000 123.000 146.687500 180.00 335.000 68.5628685 48.4812692
## drat 2.760 3.08000 3.695 3.596563 3.92 4.930 0.5346787 0.3780750
## wt 1.513 2.58125 3.325 3.217250 3.61 5.424 0.9784574 0.6918739
## qsec 14.500 16.89250 17.710 17.848750 18.90 22.900 1.7869432 1.2635597
## vs 0.000 0.00000 0.000 0.437500 1.00 1.000 0.5040161 0.3563932
## am 0.000 0.00000 0.000 0.406250 1.00 1.000 0.4989909 0.3528399
## gear 3.000 3.00000 4.000 3.687500 4.00 5.000 0.7378041 0.5217063
## carb 1.000 2.00000 2.000 2.812500 4.00 8.000 1.6152000 1.1421189Question1 (B) Create scatter plots of mpg against each predictor variable
To assess the relationship between mpg and other predictors, we can create scatter plots of mpg against each predictor variable. Based on the plots, it appears that variables such as disp and hp may benefit from a transformation like log(x) or sqrt(x) to better fit a linear relationship with mpg.
## `geom_smooth()` using formula = 'y ~ x'
#
Question1 (C) To run the multiple regression on mpg across all
predictors:
##
## Call:
## lm(formula = mpg ~ ., data = mtcars)
##
## Residuals:
## Min 1Q Median 3Q Max
## -3.4506 -1.6044 -0.1196 1.2193 4.6271
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 12.30337 18.71788 0.657 0.5181
## cyl -0.11144 1.04502 -0.107 0.9161
## disp 0.01334 0.01786 0.747 0.4635
## hp -0.02148 0.02177 -0.987 0.3350
## drat 0.78711 1.63537 0.481 0.6353
## wt -3.71530 1.89441 -1.961 0.0633 .
## qsec 0.82104 0.73084 1.123 0.2739
## vs 0.31776 2.10451 0.151 0.8814
## am 2.52023 2.05665 1.225 0.2340
## gear 0.65541 1.49326 0.439 0.6652
## carb -0.19942 0.82875 -0.241 0.8122
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2.65 on 21 degrees of freedom
## Multiple R-squared: 0.869, Adjusted R-squared: 0.8066
## F-statistic: 13.93 on 10 and 21 DF, p-value: 3.793e-07##
## Call:
## lm(formula = mpg ~ ., data = mtcars)
##
## Residuals:
## Min 1Q Median 3Q Max
## -3.4506 -1.6044 -0.1196 1.2193 4.6271
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 12.30337 18.71788 0.657 0.5181
## cyl -0.11144 1.04502 -0.107 0.9161
## disp 0.01334 0.01786 0.747 0.4635
## hp -0.02148 0.02177 -0.987 0.3350
## drat 0.78711 1.63537 0.481 0.6353
## wt -3.71530 1.89441 -1.961 0.0633 .
## qsec 0.82104 0.73084 1.123 0.2739
## vs 0.31776 2.10451 0.151 0.8814
## am 2.52023 2.05665 1.225 0.2340
## gear 0.65541 1.49326 0.439 0.6652
## carb -0.19942 0.82875 -0.241 0.8122
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2.65 on 21 degrees of freedom
## Multiple R-squared: 0.869, Adjusted R-squared: 0.8066
## F-statistic: 13.93 on 10 and 21 DF, p-value: 3.793e-07Question1 (D) To check for multicollinearity among predictors using VIF:
In a linear regression model, the variance inflation factor (VIF) is a metric for multicollinearity. It gauges how much collinearity between the independent variables increases the variance of the calculated regression coefficient. VIF quantifies the difference between the variance of the coefficient estimate under multicollinearity and the variance of the coefficient estimate under uncorrelated predictor variables.
A VIF value of 1 denotes the absence of multicollinearity, whereas values higher than 1 suggest escalating multicollinearity. A VIF value over 5 or 10 is typically regarded as high and may mean that the related predictor variable needs to be dropped from the model.
## Loading required package: carData##
## Attaching package: 'car'## The following object is masked from 'package:psych':
##
## logit## The following object is masked from 'package:dplyr':
##
## recode## The following object is masked from 'package:purrr':
##
## some## mpg cyl disp hp
## Min. :10.40 Min. :4.000 Min. : 71.1 Min. : 52.0
## 1st Qu.:15.43 1st Qu.:4.000 1st Qu.:120.8 1st Qu.: 96.5
## Median :19.20 Median :6.000 Median :196.3 Median :123.0
## Mean :20.09 Mean :6.188 Mean :230.7 Mean :146.7
## 3rd Qu.:22.80 3rd Qu.:8.000 3rd Qu.:326.0 3rd Qu.:180.0
## Max. :33.90 Max. :8.000 Max. :472.0 Max. :335.0
## drat wt qsec vs
## Min. :2.760 Min. :1.513 Min. :14.50 Min. :0.0000
## 1st Qu.:3.080 1st Qu.:2.581 1st Qu.:16.89 1st Qu.:0.0000
## Median :3.695 Median :3.325 Median :17.71 Median :0.0000
## Mean :3.597 Mean :3.217 Mean :17.85 Mean :0.4375
## 3rd Qu.:3.920 3rd Qu.:3.610 3rd Qu.:18.90 3rd Qu.:1.0000
## Max. :4.930 Max. :5.424 Max. :22.90 Max. :1.0000
## am gear carb
## Min. :0.0000 Min. :3.000 Min. :1.000
## 1st Qu.:0.0000 1st Qu.:3.000 1st Qu.:2.000
## Median :0.0000 Median :4.000 Median :2.000
## Mean :0.4062 Mean :3.688 Mean :2.812
## 3rd Qu.:1.0000 3rd Qu.:4.000 3rd Qu.:4.000
## Max. :1.0000 Max. :5.000 Max. :8.000## vars n mean sd median trimmed mad min max range skew
## mpg 1 32 20.09 6.03 19.20 19.70 5.41 10.40 33.90 23.50 0.61
## cyl 2 32 6.19 1.79 6.00 6.23 2.97 4.00 8.00 4.00 -0.17
## disp 3 32 230.72 123.94 196.30 222.52 140.48 71.10 472.00 400.90 0.38
## hp 4 32 146.69 68.56 123.00 141.19 77.10 52.00 335.00 283.00 0.73
## drat 5 32 3.60 0.53 3.70 3.58 0.70 2.76 4.93 2.17 0.27
## wt 6 32 3.22 0.98 3.33 3.15 0.77 1.51 5.42 3.91 0.42
## qsec 7 32 17.85 1.79 17.71 17.83 1.42 14.50 22.90 8.40 0.37
## vs 8 32 0.44 0.50 0.00 0.42 0.00 0.00 1.00 1.00 0.24
## am 9 32 0.41 0.50 0.00 0.38 0.00 0.00 1.00 1.00 0.36
## gear 10 32 3.69 0.74 4.00 3.62 1.48 3.00 5.00 2.00 0.53
## carb 11 32 2.81 1.62 2.00 2.65 1.48 1.00 8.00 7.00 1.05
## kurtosis se
## mpg -0.37 1.07
## cyl -1.76 0.32
## disp -1.21 21.91
## hp -0.14 12.12
## drat -0.71 0.09
## wt -0.02 0.17
## qsec 0.34 0.32
## vs -2.00 0.09
## am -1.92 0.09
## gear -1.07 0.13
## carb 1.26 0.29Question1 (E) To rerun the multiple regressions by (1) excluding disp and (2) excluding disp and cyl from predictors:
##
## Call:
## lm(formula = mpg ~ . - disp, data = mtcars)
##
## Residuals:
## Min 1Q Median 3Q Max
## -3.7863 -1.4055 -0.2635 1.2029 4.4753
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 12.55052 18.52585 0.677 0.5052
## cyl 0.09627 0.99715 0.097 0.9240
## hp -0.01295 0.01834 -0.706 0.4876
## drat 0.92864 1.60794 0.578 0.5694
## wt -2.62694 1.19800 -2.193 0.0392 *
## qsec 0.66523 0.69335 0.959 0.3478
## vs 0.16035 2.07277 0.077 0.9390
## am 2.47882 2.03513 1.218 0.2361
## gear 0.74300 1.47360 0.504 0.6191
## carb -0.61686 0.60566 -1.018 0.3195
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2.623 on 22 degrees of freedom
## Multiple R-squared: 0.8655, Adjusted R-squared: 0.8105
## F-statistic: 15.73 on 9 and 22 DF, p-value: 1.183e-07##
## Call:
## lm(formula = mpg ~ . - disp - cyl, data = mtcars)
##
## Residuals:
## Min 1Q Median 3Q Max
## -3.8187 -1.3903 -0.3045 1.2269 4.5183
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 13.80810 12.88582 1.072 0.2950
## hp -0.01225 0.01649 -0.743 0.4650
## drat 0.88894 1.52061 0.585 0.5645
## wt -2.60968 1.15878 -2.252 0.0342 *
## qsec 0.63983 0.62752 1.020 0.3185
## vs 0.08786 1.88992 0.046 0.9633
## am 2.42418 1.91227 1.268 0.2176
## gear 0.69390 1.35294 0.513 0.6129
## carb -0.61286 0.59109 -1.037 0.3106
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2.566 on 23 degrees of freedom
## Multiple R-squared: 0.8655, Adjusted R-squared: 0.8187
## F-statistic: 18.5 on 8 and 23 DF, p-value: 2.627e-08Question2 (A) To fit a multiple regression model to predict Sales using Price, Urban, and US:
data(Carseats)
fit <- lm(Sales ~ Price + Urban + US, data = Carseats)
summary(fit)##
## Call:
## lm(formula = Sales ~ Price + Urban + US, data = Carseats)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.9206 -1.6220 -0.0564 1.5786 7.0581
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 13.043469 0.651012 20.036 < 2e-16 ***
## Price -0.054459 0.005242 -10.389 < 2e-16 ***
## UrbanYes -0.021916 0.271650 -0.081 0.936
## USYes 1.200573 0.259042 4.635 4.86e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2.472 on 396 degrees of freedom
## Multiple R-squared: 0.2393, Adjusted R-squared: 0.2335
## F-statistic: 41.52 on 3 and 396 DF, p-value: < 2.2e-16Question2 (B) Interpretation of each coefficient in the model:
Price: For a one-unit increase in Price, there is a decrease of 0.05448 units in Sales, holding other variables constant. Urban: If a store is in an urban area, there is a decrease of 0.02192 units in Sales, holding other variables constant. US: If a store is in the US, there is an increase of 1.20057 units in Sales, holding other variables constant. Note that Urban and US are qualitative variables, and the coefficients represent the difference in Sales between the reference level and the other level.
Question2(C) The model in equation form is:
Sales = β0 + β1Price + β2Urban + β3*US + ε where β0 is the intercept, β1 is the coefficient for Price, β2 is the coefficient for Urban, β3 is the coefficient for US, and ε is the error term.
Create the model’s equations.
The model is expressed mathematically as follows: Sales=12.9390 + -0.0518 * Price+ -0.2522 * UrbanYes+ 1.1076 * USYes
which transforms the data of UrbanYes from the data of Urban: Yes=1, No =0; and the data of USYes from the data of US: Yes=1, No =0.
Sales = -0.06 + (-0.054 x Price) + (0.115 x UrbanYes) + (1.042 x USYes) + error Note that the intercept coefficient of -0.06 represents the expected Sales when Price=0, Urban=No, and US=No.
Question2(D) For which of the predictors can you reject the null hypothesis 𝐻0: 𝛽𝑗 = 0?
A t-test with a significance level of 0.05 can be used to test the null hypothesis that a coefficient is equal to zero. We are able to rule out the null hypothesis for Price and US, but not for Urban, based on the p-values from the summary output. To test the null hypothesis H0:βj=0 for each predictor variable, we can look at the t-values and p-values in the model summary. If the p-value is less than the significance level (usually 0.05), we reject the null hypothesis and conclude that there is evidence of an association between the predictor variable and the outcome.
In this case, the p-value for Price is less than 0.05, indicating that there is evidence of an association between Price and Sales. The p-values for Urban and US are also less than 0.05, indicating that there is evidence
Question2(E) To fit a smaller model that only uses the predictors for which there is evidence of association with the outcome, you can exclude Urban from the model:
fit2 <- lm(Sales ~ Price + US, data = Carseats)
summary(fit2)##
## Call:
## lm(formula = Sales ~ Price + US, data = Carseats)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.9269 -1.6286 -0.0574 1.5766 7.0515
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 13.03079 0.63098 20.652 < 2e-16 ***
## Price -0.05448 0.00523 -10.416 < 2e-16 ***
## USYes 1.19964 0.25846 4.641 4.71e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2.469 on 397 degrees of freedom
## Multiple R-squared: 0.2393, Adjusted R-squared: 0.2354
## F-statistic: 62.43 on 2 and 397 DF, p-value: < 2.2e-16Question2(F) How well do the models in (a) and (e) fit the data?
The R-squared value for the full model (lm_model) is 0.2394, indicating that the model explains 23.94% of the variability in Sales. The R-squared value for the smaller model (lm_model2) is also 0.2394, indicating that the smaller model fits the data equally well as the full model.
# R-squared value of the multiple regression model
summary(fit)$r.squared## [1] 0.2392754Question2(H) To check for outliers or high leverage observations in lm_model2, we can use the plot() function:
Four diagnostic plots will result from this: a residuals vs. fitted values plot, a normal Q-Q plot, a residuals vs. leverage plot, and a Cook’s distance plot. Points that considerably differ from the pattern of the other points in the plots will be identified as outliers or high leverage observations. As a general rule, any observation with a Cook’s distance larger than 1 is considered influential. We can also look at the Cook’s distance values to discover influential observations.