Predicting the duration of car keep in inventory before sell
Load the packages
pacman::p_load(tidyverse, caret, corrplot, broom, ggpubr, MASS, relaimpo, car, e1071, interplot)Read the file and explore the dataset
setwd("C:/Users/ngsook/Desktop/NUS EBA/Semester 2/Predictive Analytic/EBA Predictive WK 2/data")
sedan <- read.csv("drive_time_sedans.csv")
dim(sedan)## [1] 17506 13head(sedan)## data.set total.cost lot.sale.days overage mileage vehicle.type
## 1 TRAIN 4037 135 YES 67341 FAMILY.LARGE
## 2 TRAIN 4662 18 NO 69384 FAMILY.SMALL
## 3 TRAIN 4459 65 NO 58239 ECONOMY
## 4 TRAIN 4279 1 NO 58999 ECONOMY
## 5 TRAIN 4472 37 NO 47234 FAMILY.MEDIUM
## 6 TRAIN 5256 12 NO 64674 FAMILY.SMALL
## domestic.import vehicle.age vehicle.age.group color.set makex state
## 1 Domestic 8 SEVEN+ SILVER BUICK CA
## 2 Import 4 FOUR SILVER MAZDA TX
## 3 Import 4 FOUR RED GEO FL
## 4 Import 3 ONE-THREE RED KIA TX
## 5 Domestic 6 SIX BLUE MERCURY NM
## 6 Domestic 3 ONE-THREE GOLD CHEVROLET CA
## make.model
## 1 BUICK.CENTURY
## 2 MAZDA.PROTEGE
## 3 GEO.METRO
## 4 KIA.SEPHIA
## 5 MERCURY.MYSTIQUE
## 6 CHEVROLET.CAVALIERsummary(sedan)## data.set total.cost lot.sale.days overage
## TEST :4376 Min. :1375 Min. : 0.00 NO :14011
## TRAIN :8753 1st Qu.:4300 1st Qu.: 7.00 YES: 3495
## VALIDATE:4377 Median :4852 Median : 24.00
## Mean :4895 Mean : 46.82
## 3rd Qu.:5444 3rd Qu.: 74.00
## Max. :9813 Max. :300.00
##
## mileage vehicle.type domestic.import vehicle.age
## Min. : 82 ECONOMY :2142 Domestic:12035 Min. : 1.000
## 1st Qu.: 67242 FAMILY.LARGE :2940 Import : 5471 1st Qu.: 4.000
## Median : 77961 FAMILY.MEDIUM:6897 Median : 5.000
## Mean : 76950 FAMILY.SMALL :3069 Mean : 5.441
## 3rd Qu.: 89090 LUXURY :2458 3rd Qu.: 7.000
## Max. :119250 Max. :12.000
##
## vehicle.age.group color.set makex state
## FIVE :4012 WHITE :3603 FORD :2287 TX :4348
## FOUR :3393 GREEN :3388 CHEVROLET :2142 CA :3621
## ONE-THREE:2127 BLUE :2509 BUICK :1674 FL :3147
## SEVEN+ :4537 SILVER :1853 NISSAN :1610 AZ :2335
## SIX :3437 PURPLE :1784 PONTIAC :1547 GA :1725
## GOLD :1719 OLDSMOBILE:1512 VA :1294
## (Other):2650 (Other) :6734 (Other):1036
## make.model
## CHEVROLET.LUMINA: 997
## FORD.TAURUS : 993
## NISSAN.SENTRA : 901
## PONTIAC.GRAND AM: 683
## FORD.ESCORT : 619
## NISSAN.ALTIMA : 584
## (Other) :12729str(sedan)## 'data.frame': 17506 obs. of 13 variables:
## $ data.set : Factor w/ 3 levels "TEST","TRAIN",..: 2 2 2 2 2 2 2 2 2 2 ...
## $ total.cost : int 4037 4662 4459 4279 4472 5256 5522 3194 4332 4249 ...
## $ lot.sale.days : int 135 18 65 1 37 12 7 16 7 42 ...
## $ overage : Factor w/ 2 levels "NO","YES": 2 1 1 1 1 1 1 1 1 1 ...
## $ mileage : int 67341 69384 58239 58999 47234 64674 92879 67197 98403 87183 ...
## $ vehicle.type : Factor w/ 5 levels "ECONOMY","FAMILY.LARGE",..: 2 4 1 1 3 4 2 4 3 3 ...
## $ domestic.import : Factor w/ 2 levels "Domestic","Import": 1 2 2 2 1 1 1 1 1 1 ...
## $ vehicle.age : int 8 4 4 3 6 3 5 6 4 6 ...
## $ vehicle.age.group: Factor w/ 5 levels "FIVE","FOUR",..: 4 2 2 3 5 3 1 5 2 5 ...
## $ color.set : Factor w/ 8 levels "BLACK","BLUE",..: 7 7 6 6 2 3 6 7 8 7 ...
## $ makex : Factor w/ 20 levels "BUICK","CADILLAC",..: 1 12 8 11 13 3 1 18 3 19 ...
## $ state : Factor w/ 8 levels "AZ","CA","FL",..: 2 7 3 7 5 2 6 7 5 3 ...
## $ make.model : Factor w/ 110 levels "ACURA.INTEGRA",..: 3 68 50 62 70 18 5 93 21 97 ...Set vehicle.age.group and vehicle.age to ordered factor
unique(sedan$vehicle.age.group)## [1] SEVEN+ FOUR ONE-THREE SIX FIVE
## Levels: FIVE FOUR ONE-THREE SEVEN+ SIXsedan$vehicle.age.group = factor(sedan$vehicle.age.group, levels = c("ONE-THREE", "FOUR", "FIVE", "SIX", "SEVEN+"), ordered = TRUE)
str(sedan)## 'data.frame': 17506 obs. of 13 variables:
## $ data.set : Factor w/ 3 levels "TEST","TRAIN",..: 2 2 2 2 2 2 2 2 2 2 ...
## $ total.cost : int 4037 4662 4459 4279 4472 5256 5522 3194 4332 4249 ...
## $ lot.sale.days : int 135 18 65 1 37 12 7 16 7 42 ...
## $ overage : Factor w/ 2 levels "NO","YES": 2 1 1 1 1 1 1 1 1 1 ...
## $ mileage : int 67341 69384 58239 58999 47234 64674 92879 67197 98403 87183 ...
## $ vehicle.type : Factor w/ 5 levels "ECONOMY","FAMILY.LARGE",..: 2 4 1 1 3 4 2 4 3 3 ...
## $ domestic.import : Factor w/ 2 levels "Domestic","Import": 1 2 2 2 1 1 1 1 1 1 ...
## $ vehicle.age : int 8 4 4 3 6 3 5 6 4 6 ...
## $ vehicle.age.group: Ord.factor w/ 5 levels "ONE-THREE"<"FOUR"<..: 5 2 2 1 4 1 3 4 2 4 ...
## $ color.set : Factor w/ 8 levels "BLACK","BLUE",..: 7 7 6 6 2 3 6 7 8 7 ...
## $ makex : Factor w/ 20 levels "BUICK","CADILLAC",..: 1 12 8 11 13 3 1 18 3 19 ...
## $ state : Factor w/ 8 levels "AZ","CA","FL",..: 2 7 3 7 5 2 6 7 5 3 ...
## $ make.model : Factor w/ 110 levels "ACURA.INTEGRA",..: 3 68 50 62 70 18 5 93 21 97 ...head(sedan)## data.set total.cost lot.sale.days overage mileage vehicle.type
## 1 TRAIN 4037 135 YES 67341 FAMILY.LARGE
## 2 TRAIN 4662 18 NO 69384 FAMILY.SMALL
## 3 TRAIN 4459 65 NO 58239 ECONOMY
## 4 TRAIN 4279 1 NO 58999 ECONOMY
## 5 TRAIN 4472 37 NO 47234 FAMILY.MEDIUM
## 6 TRAIN 5256 12 NO 64674 FAMILY.SMALL
## domestic.import vehicle.age vehicle.age.group color.set makex state
## 1 Domestic 8 SEVEN+ SILVER BUICK CA
## 2 Import 4 FOUR SILVER MAZDA TX
## 3 Import 4 FOUR RED GEO FL
## 4 Import 3 ONE-THREE RED KIA TX
## 5 Domestic 6 SIX BLUE MERCURY NM
## 6 Domestic 3 ONE-THREE GOLD CHEVROLET CA
## make.model
## 1 BUICK.CENTURY
## 2 MAZDA.PROTEGE
## 3 GEO.METRO
## 4 KIA.SEPHIA
## 5 MERCURY.MYSTIQUE
## 6 CHEVROLET.CAVALIERRemove overage because this is category for dependent variable lot.sale.days.
Not going to predit lot.sale.days by using the overage category
sedan = sedan[,-c(4)]
head(sedan)## data.set total.cost lot.sale.days mileage vehicle.type domestic.import
## 1 TRAIN 4037 135 67341 FAMILY.LARGE Domestic
## 2 TRAIN 4662 18 69384 FAMILY.SMALL Import
## 3 TRAIN 4459 65 58239 ECONOMY Import
## 4 TRAIN 4279 1 58999 ECONOMY Import
## 5 TRAIN 4472 37 47234 FAMILY.MEDIUM Domestic
## 6 TRAIN 5256 12 64674 FAMILY.SMALL Domestic
## vehicle.age vehicle.age.group color.set makex state
## 1 8 SEVEN+ SILVER BUICK CA
## 2 4 FOUR SILVER MAZDA TX
## 3 4 FOUR RED GEO FL
## 4 3 ONE-THREE RED KIA TX
## 5 6 SIX BLUE MERCURY NM
## 6 3 ONE-THREE GOLD CHEVROLET CA
## make.model
## 1 BUICK.CENTURY
## 2 MAZDA.PROTEGE
## 3 GEO.METRO
## 4 KIA.SEPHIA
## 5 MERCURY.MYSTIQUE
## 6 CHEVROLET.CAVALIERHow the target variable (lot.sale.days) distributed?
Try to transform the target variable with square root to look symmetry
p1 = sedan %>%
ggplot(aes(y=lot.sale.days)) + geom_boxplot()
p2 = sedan %>%
ggplot(aes(lot.sale.days)) + geom_histogram()
p3 = sedan %>%
ggplot(aes(sqrt(lot.sale.days))) + geom_histogram()
p4 = sedan %>%
ggplot(aes(y=sqrt(lot.sale.days))) + geom_boxplot()
ggarrange(p1,p2,p3,p4 + rremove("x.text"),
labels = c("A", "B" ,"C" ,"D"),
ncol = 2, nrow = 2)
Verify the skewness of target variable
Skewness (symmetry) can be improve by taking sqrt of lot.sale.days
skewness(sedan$lot.sale.days)## [1] 1.534117skewness(sqrt(sedan$lot.sale.days))## [1] 0.5840478The outlier will affect the skewness of the data??
Outlier value
outliers = sedan %>%
dplyr::filter((abs(lot.sale.days - median(lot.sale.days)) > 2*sd(lot.sale.days)))
dim(sedan)## [1] 17506 12dim(outliers)## [1] 1489 12sedan_no_outlier <- sedan %>%
dplyr::filter(!abs(lot.sale.days - median(lot.sale.days)) > 2*sd(lot.sale.days))
dim(sedan_no_outlier)## [1] 16017 12colnames(sedan_no_outlier)## [1] "data.set" "total.cost" "lot.sale.days"
## [4] "mileage" "vehicle.type" "domestic.import"
## [7] "vehicle.age" "vehicle.age.group" "color.set"
## [10] "makex" "state" "make.model"How the target variable (lot.sale.days) distributed after remove the outlier?
p1 = sedan_no_outlier %>%
ggplot(aes(y=lot.sale.days)) + geom_boxplot()
p2 = sedan_no_outlier %>%
ggplot(aes(lot.sale.days)) + geom_histogram(binwidth = 5)
p3 = sedan_no_outlier %>%
ggplot(aes(sqrt(lot.sale.days))) + geom_histogram(binwidth = 0.5)
ggarrange(p1,p2, p3,
labels = c("A", "B", "C"),
ncol =2, nrow =2)
Correlation plots
Correlation between mileage, total.cost and vehicle age with lot.sale.days is very weak
options(repr.plot.width = 4, repr.plot.height = 3)
corrplot::corrplot(cor(sedan_no_outlier[, sapply(sedan_no_outlier, is.numeric)],
use = "complete.obs"), method = "number", type ="lower")
Test the linearity with the continuous independent variables
Variables: total.cost, mileage, vehicle.age
p1 = sedan_no_outlier %>%
ggplot(aes(x= total.cost, y=lot.sale.days)) + geom_point() + geom_smooth(method = "loess")
p2 = sedan_no_outlier %>%
ggplot(aes(x= mileage, y=lot.sale.days)) + geom_point() + geom_smooth(method = "loess")
p3 = sedan_no_outlier %>%
ggplot(aes(x= vehicle.age, y= lot.sale.days)) + geom_point() + geom_smooth(method = 'loess')
ggarrange(p1,p2, p3 + rremove("x.text"),
labels = c("A", "B", "C"),
ncol = 2, nrow = 2)
Transformation
options(repr.plot.width = 8, repr.plot.height = 4)
p1 = sedan_no_outlier %>%
ggplot(aes(x= sqrt(total.cost), y=lot.sale.days)) + geom_point() + geom_smooth(method = "loess")
p2 = sedan_no_outlier %>%
ggplot(aes(x= sqrt(mileage), y=lot.sale.days)) + geom_point() + geom_smooth(method = "loess")
p3 = sedan_no_outlier %>%
ggplot(aes(x=sqrt(vehicle.age), y= lot.sale.days)) + geom_point() + geom_smooth(method = 'loess')
ggarrange(p1,p2, p3 + rremove("x.text"),
labels = c("A", "B", "C"),
ncol = 2, nrow = 2)
Put the transformation into the sedan dataset
cols_old = c('total.cost', 'mileage')
cols = c('s_total.cost', 's_mileage')
sedan_no_outlier[,cols] = sapply(sedan_no_outlier[cols_old], sqrt)
head(sedan_no_outlier)## data.set total.cost lot.sale.days mileage vehicle.type domestic.import
## 1 TRAIN 4662 18 69384 FAMILY.SMALL Import
## 2 TRAIN 4459 65 58239 ECONOMY Import
## 3 TRAIN 4279 1 58999 ECONOMY Import
## 4 TRAIN 4472 37 47234 FAMILY.MEDIUM Domestic
## 5 TRAIN 5256 12 64674 FAMILY.SMALL Domestic
## 6 TRAIN 5522 7 92879 FAMILY.LARGE Domestic
## vehicle.age vehicle.age.group color.set makex state
## 1 4 FOUR SILVER MAZDA TX
## 2 4 FOUR RED GEO FL
## 3 3 ONE-THREE RED KIA TX
## 4 6 SIX BLUE MERCURY NM
## 5 3 ONE-THREE GOLD CHEVROLET CA
## 6 5 FIVE RED BUICK NV
## make.model s_total.cost s_mileage
## 1 MAZDA.PROTEGE 68.27884 263.4084
## 2 GEO.METRO 66.77574 241.3276
## 3 KIA.SEPHIA 65.41407 242.8971
## 4 MERCURY.MYSTIQUE 66.87301 217.3338
## 5 CHEVROLET.CAVALIER 72.49828 254.3108
## 6 BUICK.LESABRE 74.31016 304.7606corrplot:: corrplot(cor(sedan_no_outlier[,sapply(sedan_no_outlier,is.numeric)]),
use = 'complete.obs', method = 'number', type = 'lower')
Initial the first model without transform variables
model1 = lm(lot.sale.days~. -s_total.cost -s_mileage, data = sedan_no_outlier)
summary(model1)##
## Call:
## lm(formula = lot.sale.days ~ . - s_total.cost - s_mileage, data = sedan_no_outlier)
##
## Residuals:
## Min 1Q Median 3Q Max
## -88.812 -22.518 -7.502 18.222 106.601
##
## Coefficients: (19 not defined because of singularities)
## Estimate Std. Error t value
## (Intercept) -7.014e+01 5.180e+00 -13.539
## data.setTRAIN 4.099e-01 6.215e-01 0.659
## data.setVALIDATE 2.508e-01 7.171e-01 0.350
## total.cost 1.761e-02 3.836e-04 45.914
## mileage 2.666e-04 1.987e-05 13.416
## vehicle.typeFAMILY.LARGE 5.901e-02 1.532e+00 0.039
## vehicle.typeFAMILY.MEDIUM -1.062e+00 1.339e+00 -0.793
## vehicle.typeFAMILY.SMALL -1.654e+00 1.137e+00 -1.454
## vehicle.typeLUXURY 3.113e-01 1.468e+00 0.212
## domestic.importImport -1.059e+01 2.313e+01 -0.458
## vehicle.age 6.287e+00 5.941e-01 10.582
## vehicle.age.group.L 5.938e+00 2.249e+00 2.641
## vehicle.age.group.Q -1.237e+00 6.529e-01 -1.894
## vehicle.age.group.C 6.367e-01 6.104e-01 1.043
## vehicle.age.group^4 -1.389e-01 5.550e-01 -0.250
## color.setBLUE 2.745e+00 1.184e+00 2.318
## color.setGOLD 2.072e-02 1.261e+00 0.016
## color.setGREEN 2.665e+00 1.124e+00 2.372
## color.setPURPLE 2.255e+00 1.246e+00 1.809
## color.setRED 2.871e+00 1.325e+00 2.166
## color.setSILVER 4.081e-01 1.235e+00 0.331
## color.setWHITE 3.342e+00 1.123e+00 2.975
## makexCADILLAC -8.268e+01 7.511e+00 -11.007
## makexCHEVROLET -9.841e+00 5.185e+00 -1.898
## makexCHRYSLER -5.137e+01 7.135e+00 -7.200
## makexDAEWOO 4.400e+01 3.958e+01 1.112
## makexDODGE -4.700e+01 3.127e+00 -15.029
## makexFORD -2.265e+01 7.307e+00 -3.100
## makexGEO -4.892e+00 2.334e+01 -0.210
## makexHONDA -4.713e+01 2.338e+01 -2.016
## makexHYUNDAI -1.864e+01 2.368e+01 -0.787
## makexKIA 9.905e+00 2.335e+01 0.424
## makexMAZDA -8.517e+00 2.332e+01 -0.365
## makexMERCURY -4.350e+00 3.411e+00 -1.276
## makexMITSUBISHI -3.846e+00 3.961e+01 -0.097
## makexNISSAN -3.041e+01 2.623e+01 -1.159
## makexOLDSMOBILE -6.322e+01 2.279e+01 -2.773
## makexOTHER -5.705e+01 2.415e+01 -2.363
## makexPLYMOUTH -1.795e+01 3.216e+00 -5.581
## makexPONTIAC -2.510e+01 3.610e+00 -6.952
## makexTOYOTA -3.905e+00 2.338e+01 -0.167
## stateCA -8.787e+00 9.732e-01 -9.029
## stateFL -1.251e+00 9.918e-01 -1.262
## stateGA -1.613e+01 1.174e+00 -13.737
## stateNM -1.539e+01 1.682e+00 -9.151
## stateNV -7.328e+00 1.681e+00 -4.359
## stateTX -9.221e+00 9.670e-01 -9.536
## stateVA -8.065e+00 1.231e+00 -6.552
## make.modelACURA.VIGOR 1.060e+01 3.269e+01 0.324
## make.modelBUICK.CENTURY -1.479e+01 3.235e+00 -4.571
## make.modelBUICK.LASABRE -7.300e+01 3.213e+01 -2.272
## make.modelBUICK.LESABRE -3.004e+01 2.963e+00 -10.137
## make.modelBUICK.PARK AVENUE -4.812e+01 4.316e+00 -11.149
## make.modelBUICK.REGAL -2.412e+01 2.965e+00 -8.135
## make.modelBUICK.ROADMASTER -5.120e+01 4.701e+00 -10.892
## make.modelBUICK.SKYLARK NA NA NA
## make.modelCADILLAC.BROUGHAM RWD 2.414e+01 1.976e+01 1.221
## make.modelCADILLAC.CATERA 2.897e+00 1.978e+01 0.146
## make.modelCADILLAC.DEVILLE 1.616e+01 7.558e+00 2.138
## make.modelCADILLAC.FLEETWOOD -1.518e+01 2.370e+01 -0.640
## make.modelCADILLAC.FLEETWOOD RWD -3.279e-01 1.278e+01 -0.026
## make.modelCADILLAC.SEVILLE NA NA NA
## make.modelCHEVROLET.CAPRICE -3.487e+01 1.512e+01 -2.306
## make.modelCHEVROLET.CAPRICE CLASSIC -5.216e+01 6.789e+00 -7.683
## make.modelCHEVROLET.CAVALIER -5.760e+00 4.834e+00 -1.191
## make.modelCHEVROLET.CELEBRITY -5.503e+01 3.246e+01 -1.695
## make.modelCHEVROLET.CORSICA 9.809e+00 5.377e+00 1.824
## make.modelCHEVROLET.LUMINA -1.681e+01 4.823e+00 -3.485
## make.modelCHEVROLET.MALIBU -3.405e+01 5.026e+00 -6.775
## make.modelCHEVROLET.METRO 2.220e+01 6.332e+00 3.505
## make.modelCHEVROLET.PRIZM NA NA NA
## make.modelCHRYSLER.CIRRUS -2.307e+00 7.395e+00 -0.312
## make.modelCHRYSLER.CONCORDE 2.795e+00 7.143e+00 0.391
## make.modelCHRYSLER.FIFTH AVENUE 2.615e+01 2.361e+01 1.108
## make.modelCHRYSLER.FWD 2.137e+01 3.273e+01 0.653
## make.modelCHRYSLER.LEBARON 1.695e+01 3.275e+01 0.517
## make.modelCHRYSLER.LHS -4.912e+00 7.626e+00 -0.644
## make.modelCHRYSLER.NEW YORKER NA NA NA
## make.modelDAEWOO.LANOS 1.249e+01 3.229e+01 0.387
## make.modelDAEWOO.LEGANZA -4.222e+01 3.347e+01 -1.261
## make.modelDAEWOO.NUBIRA -4.271e+01 3.244e+01 -1.317
## make.modelDAEWOO.NUBIRA SX NA NA NA
## make.modelDODGE.DYNASTY 2.349e+01 2.277e+01 1.032
## make.modelDODGE.INTREPID 3.596e+00 2.677e+00 1.343
## make.modelDODGE.NEON 2.885e+01 2.832e+00 10.188
## make.modelDODGE.SPIRIT 3.173e+01 1.232e+01 2.575
## make.modelDODGE.STRATUS NA NA NA
## make.modelEAGLE.VISION 1.515e+01 2.451e+01 0.618
## make.modelEAGLE.VISION TSI -7.089e+00 4.001e+01 -0.177
## make.modelFORD.CONTOUR -4.550e+00 7.116e+00 -0.639
## make.modelFORD.CROWN VICTORIA -1.897e+01 7.436e+00 -2.551
## make.modelFORD.ESCORT 1.528e+01 7.052e+00 2.167
## make.modelFORD.FOCUS 2.439e+00 1.978e+01 0.123
## make.modelFORD.TAURUS -8.440e+00 7.040e+00 -1.199
## make.modelFORD.TEMPO NA NA NA
## make.modelGEO.METRO 2.551e+01 4.783e+00 5.334
## make.modelGEO.PRIZM NA NA NA
## make.modelHONDA.ACCORD -1.921e+01 3.396e+00 -5.659
## make.modelHONDA.CIVIC NA NA NA
## make.modelHYUNDAI.ACCENT 3.095e+01 5.605e+00 5.522
## make.modelHYUNDAI.ACCENT GL -1.241e+01 3.232e+01 -0.384
## make.modelHYUNDAI.ELANTRA 1.785e+01 5.198e+00 3.434
## make.modelHYUNDAI.SONATA NA NA NA
## make.modelINFINITI.G20 6.521e+00 7.530e+00 0.866
## make.modelINFINITI.I30 -1.909e+00 1.728e+01 -0.110
## make.modelINFINITI.J30 -2.371e+01 8.987e+00 -2.638
## make.modelINFINITI.Q45 -2.430e+01 1.371e+01 -1.773
## make.modelKIA.SEPHIA NA NA NA
## make.modelLEXUS.300 -4.678e+01 1.957e+01 -2.390
## make.modelLINCOLN.TOWN CAR -4.948e+00 2.430e+01 -0.204
## make.modelMAZDA.626 -3.106e+01 2.547e+00 -12.194
## make.modelMAZDA.929 -4.895e+01 3.208e+01 -1.526
## make.modelMAZDA.MILLENIA -5.886e+01 9.879e+00 -5.959
## make.modelMAZDA.PROTEGE NA NA NA
## make.modelMERCURY.GRAND MARQUIS -4.569e+01 3.775e+00 -12.105
## make.modelMERCURY.MYSTIQUE -2.366e+01 3.851e+00 -6.145
## make.modelMERCURY.SABLE -2.303e+01 3.540e+00 -6.506
## make.modelMERCURY.TOPAZ -1.984e+01 1.243e+01 -1.596
## make.modelMERCURY.TRACER NA NA NA
## make.modelMITSUBISHI.DIAMANTE -6.176e+00 3.928e+01 -0.157
## make.modelMITSUBISHI.GALANT -2.622e+01 3.212e+01 -0.816
## make.modelMITSUBISHI.MIRAGE -5.314e+00 3.218e+01 -0.165
## make.modelMITSUBISHI.MIRAGE DE NA NA NA
## make.modelNISSAN.ALTIMA -1.009e+01 1.224e+01 -0.824
## make.modelNISSAN.MAXIMA -2.741e+01 1.253e+01 -2.188
## make.modelNISSAN.SENTRA 2.335e+01 1.220e+01 1.914
## make.modelNISSAN.STANZA NA NA NA
## make.modelOLDSMOBILE.88 3.491e+01 2.271e+01 1.537
## make.modelOLDSMOBILE.98 2.362e+01 2.292e+01 1.031
## make.modelOLDSMOBILE.ACHIEVA 6.186e+01 2.279e+01 2.715
## make.modelOLDSMOBILE.ALERO 1.669e+01 2.424e+01 0.688
## make.modelOLDSMOBILE.AURORA -7.676e+00 2.389e+01 -0.321
## make.modelOLDSMOBILE.CALAIS 1.477e+01 3.210e+01 0.460
## make.modelOLDSMOBILE.CIERA 6.319e+01 2.281e+01 2.770
## make.modelOLDSMOBILE.CUTLASS 2.183e+01 2.314e+01 0.943
## make.modelOLDSMOBILE.CUTLASS SUPREME 4.512e+01 2.272e+01 1.986
## make.modelOLDSMOBILE.INTRIGUE 9.240e+00 2.334e+01 0.396
## make.modelOLDSMOBILE.REGENCY NA NA NA
## make.modelPLYMOUTH.ACCLAIM 2.644e+00 1.328e+01 0.199
## make.modelPLYMOUTH.BREEZE -1.511e+01 3.263e+00 -4.631
## make.modelPLYMOUTH.NEON NA NA NA
## make.modelPONTIAC.BONNEVILLE -1.824e+01 3.351e+00 -5.442
## make.modelPONTIAC.GRAND AM -6.874e+00 3.043e+00 -2.259
## make.modelPONTIAC.GRAND PRIX -6.449e+00 3.501e+00 -1.842
## make.modelPONTIAC.SUNBIRD -7.597e+00 1.462e+01 -0.520
## make.modelPONTIAC.SUNFIRE NA NA NA
## make.modelSATURN.S-SERIES -8.171e+00 4.006e+01 -0.204
## make.modelSATURN.SL 6.724e+01 4.008e+01 1.678
## make.modelSATURN.SL-1 1.130e+01 4.002e+01 0.282
## make.modelSUZUKI.ESTEEM 6.568e+01 7.874e+00 8.342
## make.modelTOYOTA.AVALON -6.165e+01 1.628e+01 -3.786
## make.modelTOYOTA.CAMRY -5.896e+01 3.309e+00 -17.817
## make.modelTOYOTA.COROLLA -1.817e+01 3.118e+00 -5.828
## make.modelTOYOTA.TERCEL NA NA NA
## Pr(>|t|)
## (Intercept) < 2e-16 ***
## data.setTRAIN 0.509585
## data.setVALIDATE 0.726490
## total.cost < 2e-16 ***
## mileage < 2e-16 ***
## vehicle.typeFAMILY.LARGE 0.969269
## vehicle.typeFAMILY.MEDIUM 0.427673
## vehicle.typeFAMILY.SMALL 0.145854
## vehicle.typeLUXURY 0.832008
## domestic.importImport 0.647207
## vehicle.age < 2e-16 ***
## vehicle.age.group.L 0.008273 **
## vehicle.age.group.Q 0.058210 .
## vehicle.age.group.C 0.296930
## vehicle.age.group^4 0.802397
## color.setBLUE 0.020456 *
## color.setGOLD 0.986886
## color.setGREEN 0.017723 *
## color.setPURPLE 0.070396 .
## color.setRED 0.030298 *
## color.setSILVER 0.741022
## color.setWHITE 0.002931 **
## makexCADILLAC < 2e-16 ***
## makexCHEVROLET 0.057690 .
## makexCHRYSLER 6.30e-13 ***
## makexDAEWOO 0.266339
## makexDODGE < 2e-16 ***
## makexFORD 0.001939 **
## makexGEO 0.833961
## makexHONDA 0.043834 *
## makexHYUNDAI 0.431199
## makexKIA 0.671410
## makexMAZDA 0.714963
## makexMERCURY 0.202125
## makexMITSUBISHI 0.922639
## makexNISSAN 0.246356
## makexOLDSMOBILE 0.005552 **
## makexOTHER 0.018155 *
## makexPLYMOUTH 2.42e-08 ***
## makexPONTIAC 3.75e-12 ***
## makexTOYOTA 0.867385
## stateCA < 2e-16 ***
## stateFL 0.207045
## stateGA < 2e-16 ***
## stateNM < 2e-16 ***
## stateNV 1.31e-05 ***
## stateTX < 2e-16 ***
## stateVA 5.85e-11 ***
## make.modelACURA.VIGOR 0.745788
## make.modelBUICK.CENTURY 4.89e-06 ***
## make.modelBUICK.LASABRE 0.023086 *
## make.modelBUICK.LESABRE < 2e-16 ***
## make.modelBUICK.PARK AVENUE < 2e-16 ***
## make.modelBUICK.REGAL 4.43e-16 ***
## make.modelBUICK.ROADMASTER < 2e-16 ***
## make.modelBUICK.SKYLARK NA
## make.modelCADILLAC.BROUGHAM RWD 0.221963
## make.modelCADILLAC.CATERA 0.883552
## make.modelCADILLAC.DEVILLE 0.032506 *
## make.modelCADILLAC.FLEETWOOD 0.521984
## make.modelCADILLAC.FLEETWOOD RWD 0.979530
## make.modelCADILLAC.SEVILLE NA
## make.modelCHEVROLET.CAPRICE 0.021113 *
## make.modelCHEVROLET.CAPRICE CLASSIC 1.64e-14 ***
## make.modelCHEVROLET.CAVALIER 0.233505
## make.modelCHEVROLET.CELEBRITY 0.090077 .
## make.modelCHEVROLET.CORSICA 0.068144 .
## make.modelCHEVROLET.LUMINA 0.000494 ***
## make.modelCHEVROLET.MALIBU 1.29e-11 ***
## make.modelCHEVROLET.METRO 0.000457 ***
## make.modelCHEVROLET.PRIZM NA
## make.modelCHRYSLER.CIRRUS 0.755021
## make.modelCHRYSLER.CONCORDE 0.695550
## make.modelCHRYSLER.FIFTH AVENUE 0.268018
## make.modelCHRYSLER.FWD 0.513899
## make.modelCHRYSLER.LEBARON 0.604884
## make.modelCHRYSLER.LHS 0.519460
## make.modelCHRYSLER.NEW YORKER NA
## make.modelDAEWOO.LANOS 0.698919
## make.modelDAEWOO.LEGANZA 0.207193
## make.modelDAEWOO.NUBIRA 0.187998
## make.modelDAEWOO.NUBIRA SX NA
## make.modelDODGE.DYNASTY 0.302262
## make.modelDODGE.INTREPID 0.179157
## make.modelDODGE.NEON < 2e-16 ***
## make.modelDODGE.SPIRIT 0.010032 *
## make.modelDODGE.STRATUS NA
## make.modelEAGLE.VISION 0.536364
## make.modelEAGLE.VISION TSI 0.859377
## make.modelFORD.CONTOUR 0.522543
## make.modelFORD.CROWN VICTORIA 0.010759 *
## make.modelFORD.ESCORT 0.030262 *
## make.modelFORD.FOCUS 0.901873
## make.modelFORD.TAURUS 0.230586
## make.modelFORD.TEMPO NA
## make.modelGEO.METRO 9.74e-08 ***
## make.modelGEO.PRIZM NA
## make.modelHONDA.ACCORD 1.55e-08 ***
## make.modelHONDA.CIVIC NA
## make.modelHYUNDAI.ACCENT 3.40e-08 ***
## make.modelHYUNDAI.ACCENT GL 0.701092
## make.modelHYUNDAI.ELANTRA 0.000596 ***
## make.modelHYUNDAI.SONATA NA
## make.modelINFINITI.G20 0.386465
## make.modelINFINITI.I30 0.912040
## make.modelINFINITI.J30 0.008350 **
## make.modelINFINITI.Q45 0.076281 .
## make.modelKIA.SEPHIA NA
## make.modelLEXUS.300 0.016857 *
## make.modelLINCOLN.TOWN CAR 0.838643
## make.modelMAZDA.626 < 2e-16 ***
## make.modelMAZDA.929 0.127083
## make.modelMAZDA.MILLENIA 2.60e-09 ***
## make.modelMAZDA.PROTEGE NA
## make.modelMERCURY.GRAND MARQUIS < 2e-16 ***
## make.modelMERCURY.MYSTIQUE 8.17e-10 ***
## make.modelMERCURY.SABLE 7.95e-11 ***
## make.modelMERCURY.TOPAZ 0.110533
## make.modelMERCURY.TRACER NA
## make.modelMITSUBISHI.DIAMANTE 0.875065
## make.modelMITSUBISHI.GALANT 0.414285
## make.modelMITSUBISHI.MIRAGE 0.868836
## make.modelMITSUBISHI.MIRAGE DE NA
## make.modelNISSAN.ALTIMA 0.409860
## make.modelNISSAN.MAXIMA 0.028675 *
## make.modelNISSAN.SENTRA 0.055627 .
## make.modelNISSAN.STANZA NA
## make.modelOLDSMOBILE.88 0.124364
## make.modelOLDSMOBILE.98 0.302760
## make.modelOLDSMOBILE.ACHIEVA 0.006644 **
## make.modelOLDSMOBILE.ALERO 0.491154
## make.modelOLDSMOBILE.AURORA 0.748008
## make.modelOLDSMOBILE.CALAIS 0.645297
## make.modelOLDSMOBILE.CIERA 0.005613 **
## make.modelOLDSMOBILE.CUTLASS 0.345613
## make.modelOLDSMOBILE.CUTLASS SUPREME 0.047021 *
## make.modelOLDSMOBILE.INTRIGUE 0.692196
## make.modelOLDSMOBILE.REGENCY NA
## make.modelPLYMOUTH.ACCLAIM 0.842175
## make.modelPLYMOUTH.BREEZE 3.67e-06 ***
## make.modelPLYMOUTH.NEON NA
## make.modelPONTIAC.BONNEVILLE 5.34e-08 ***
## make.modelPONTIAC.GRAND AM 0.023899 *
## make.modelPONTIAC.GRAND PRIX 0.065517 .
## make.modelPONTIAC.SUNBIRD 0.603392
## make.modelPONTIAC.SUNFIRE NA
## make.modelSATURN.S-SERIES 0.838371
## make.modelSATURN.SL 0.093464 .
## make.modelSATURN.SL-1 0.777689
## make.modelSUZUKI.ESTEEM < 2e-16 ***
## make.modelTOYOTA.AVALON 0.000154 ***
## make.modelTOYOTA.CAMRY < 2e-16 ***
## make.modelTOYOTA.COROLLA 5.71e-09 ***
## make.modelTOYOTA.TERCEL NA
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 31.99 on 15882 degrees of freedom
## Multiple R-squared: 0.2164, Adjusted R-squared: 0.2098
## F-statistic: 32.74 on 134 and 15882 DF, p-value: < 2.2e-16Looks like there is trend on residual plot
options(repr.plot.width = 8, repr.plot.height = 6)
par(mfrow = c(2,2))
plot(model1)
model_stat = augment(model1)
head(model_stat)## # A tibble: 6 x 21
## lot.sale.days data.set total.cost mileage vehicle.type domestic.import
## <int> <fct> <int> <int> <fct> <fct>
## 1 18 TRAIN 4662 69384 FAMILY.SMALL Import
## 2 65 TRAIN 4459 58239 ECONOMY Import
## 3 1 TRAIN 4279 58999 ECONOMY Import
## 4 37 TRAIN 4472 47234 FAMILY.MEDI~ Domestic
## 5 12 TRAIN 5256 64674 FAMILY.SMALL Domestic
## 6 7 TRAIN 5522 92879 FAMILY.LARGE Domestic
## # ... with 15 more variables: vehicle.age <int>, vehicle.age.group <ord>,
## # color.set <fct>, makex <fct>, state <fct>, make.model <fct>,
## # s_total.cost <dbl>, s_mileage <dbl>, .fitted <dbl>, .se.fit <dbl>,
## # .resid <dbl>, .hat <dbl>, .sigma <dbl>, .cooksd <dbl>,
## # .std.resid <dbl>Check the residual distribution (normal)
options(repr.plot.width = 4, repr.plot.height = 3)
model_stat %>%
ggplot(aes(.resid)) + geom_histogram()
Use the train data assigned for learning the model and fine tune
unique(sedan_no_outlier$data.set)## [1] TRAIN TEST VALIDATE
## Levels: TEST TRAIN VALIDATEtrain_set <- sedan_no_outlier %>%
dplyr::filter(data.set == "TRAIN")
dim(train_set)## [1] 7990 14test_set <- sedan_no_outlier %>%
dplyr::filter(data.set == "TEST")
dim(test_set)## [1] 4019 14validate_set <- sedan_no_outlier %>%
dplyr::filter(data.set == "VALIDATE")
dim(validate_set)## [1] 4008 14Use transformed variable to create model on train_set data
Remove makex since there is make.model showing detail information of makex + model
model2 <- lm(lot.sale.days~. -total.cost -mileage -makex, data = train_set[,-1])
summary(model2)##
## Call:
## lm(formula = lot.sale.days ~ . - total.cost - mileage - makex,
## data = train_set[, -1])
##
## Residuals:
## Min 1Q Median 3Q Max
## -81.777 -22.294 -7.307 18.446 107.425
##
## Coefficients:
## Estimate Std. Error t value
## (Intercept) -247.06088 27.05084 -9.133
## vehicle.typeFAMILY.LARGE -0.93512 2.18837 -0.427
## vehicle.typeFAMILY.MEDIUM -1.42562 1.91419 -0.745
## vehicle.typeFAMILY.SMALL -2.48812 1.63903 -1.518
## vehicle.typeLUXURY -1.08357 2.09683 -0.517
## domestic.importImport -7.19210 23.44411 -0.307
## vehicle.age 7.09803 0.83310 8.520
## vehicle.age.group.L 3.57812 3.14418 1.138
## vehicle.age.group.Q 0.14244 0.92190 0.155
## vehicle.age.group.C 0.07410 0.86101 0.086
## vehicle.age.group^4 0.94394 0.78838 1.197
## color.setBLUE 2.82072 1.64354 1.716
## color.setGOLD -0.26742 1.77751 -0.150
## color.setGREEN 2.65936 1.55646 1.709
## color.setPURPLE 1.14624 1.74361 0.657
## color.setRED 2.91384 1.86587 1.562
## color.setSILVER 0.33475 1.73459 0.193
## color.setWHITE 1.84872 1.56847 1.179
## stateCA -8.89395 1.38399 -6.426
## stateFL -2.44660 1.40777 -1.738
## stateGA -18.45235 1.66678 -11.071
## stateNM -15.98555 2.42272 -6.598
## stateNV -3.76914 2.40583 -1.567
## stateTX -10.52218 1.37109 -7.674
## stateVA -8.78736 1.74682 -5.030
## make.modelBUICK.CENTURY 49.56000 25.18978 1.967
## make.modelBUICK.LESABRE 31.68573 25.11213 1.262
## make.modelBUICK.PARK AVENUE 22.12941 25.42448 0.870
## make.modelBUICK.REGAL 36.04387 25.13182 1.434
## make.modelBUICK.ROADMASTER 14.02984 25.59121 0.548
## make.modelBUICK.SKYLARK 59.76065 25.20893 2.371
## make.modelCADILLAC.BROUGHAM RWD -27.94732 40.49617 -0.690
## make.modelCADILLAC.CATERA -15.18916 31.02326 -0.490
## make.modelCADILLAC.DEVILLE -5.79937 25.27301 -0.229
## make.modelCADILLAC.FLEETWOOD RWD -22.80571 28.79308 -0.792
## make.modelCADILLAC.SEVILLE -27.01670 26.93130 -1.003
## make.modelCHEVROLET.CAPRICE 9.00089 33.67683 0.267
## make.modelCHEVROLET.CAPRICE CLASSIC 1.67924 25.88965 0.065
## make.modelCHEVROLET.CAVALIER 47.12868 25.10255 1.877
## make.modelCHEVROLET.CELEBRITY 2.59399 40.65522 0.064
## make.modelCHEVROLET.CORSICA 65.61718 25.32069 2.591
## make.modelCHEVROLET.LUMINA 37.54507 25.05878 1.498
## make.modelCHEVROLET.MALIBU 17.51045 25.11644 0.697
## make.modelCHEVROLET.METRO 67.45583 25.83808 2.611
## make.modelCHEVROLET.PRIZM 50.85900 24.25732 2.097
## make.modelCHRYSLER.CIRRUS 4.87829 25.39394 0.192
## make.modelCHRYSLER.CONCORDE 13.85729 25.27754 0.548
## make.modelCHRYSLER.FIFTH AVENUE -20.86279 40.52627 -0.515
## make.modelCHRYSLER.LEBARON 33.96775 40.59120 0.837
## make.modelCHRYSLER.LHS 3.20527 25.48904 0.126
## make.modelCHRYSLER.NEW YORKER 9.64120 26.64154 0.362
## make.modelDAEWOO.LANOS 116.49131 11.44489 10.178
## make.modelDAEWOO.LEGANZA 32.55111 20.55022 1.584
## make.modelDAEWOO.NUBIRA 66.33663 12.03993 5.510
## make.modelDAEWOO.NUBIRA SX 103.79981 33.18951 3.127
## make.modelDODGE.DYNASTY 48.93465 40.64473 1.204
## make.modelDODGE.INTREPID 16.30870 25.09328 0.650
## make.modelDODGE.NEON 47.38876 25.16909 1.883
## make.modelDODGE.SPIRIT 54.06114 33.78382 1.600
## make.modelDODGE.STRATUS 15.83349 25.16444 0.629
## make.modelEAGLE.VISION 21.42229 25.98685 0.824
## make.modelEAGLE.VISION TSI -0.86674 40.51355 -0.021
## make.modelFORD.CONTOUR 31.97869 25.08361 1.275
## make.modelFORD.CROWN VICTORIA 19.65357 25.24170 0.779
## make.modelFORD.ESCORT 54.27242 25.15260 2.158
## make.modelFORD.FOCUS 71.15666 40.52948 1.756
## make.modelFORD.TAURUS 31.73881 25.04114 1.267
## make.modelFORD.TEMPO 49.72098 26.62605 1.867
## make.modelGEO.METRO 78.85199 10.71072 7.362
## make.modelGEO.PRIZM 54.29090 9.18737 5.909
## make.modelHONDA.ACCORD -5.68631 9.26832 -0.614
## make.modelHONDA.CIVIC 11.78751 9.21841 1.279
## make.modelHYUNDAI.ACCENT 68.16887 10.00694 6.812
## make.modelHYUNDAI.ACCENT GL 27.80544 33.12891 0.839
## make.modelHYUNDAI.ELANTRA 59.79619 9.83030 6.083
## make.modelHYUNDAI.SONATA 40.89838 10.53744 3.881
## make.modelINFINITI.G20 4.70484 10.40609 0.452
## make.modelINFINITI.I30 22.59209 24.16960 0.935
## make.modelINFINITI.J30 -23.39801 14.79950 -1.581
## make.modelINFINITI.Q45 -21.59820 20.30344 -1.064
## make.modelKIA.SEPHIA 68.19487 9.35962 7.286
## make.modelLEXUS.300 -58.82665 33.06813 -1.779
## make.modelLINCOLN.TOWN CAR -0.03379 25.49017 -0.001
## make.modelMAZDA.626 20.20911 9.07349 2.227
## make.modelMAZDA.MILLENIA -9.55314 14.84221 -0.644
## make.modelMAZDA.PROTEGE 48.55169 9.10926 5.330
## make.modelMERCURY.GRAND MARQUIS 9.62764 25.24845 0.381
## make.modelMERCURY.MYSTIQUE 35.82250 25.34394 1.413
## make.modelMERCURY.SABLE 34.13338 25.21564 1.354
## make.modelMERCURY.TOPAZ 33.28902 31.10472 1.070
## make.modelMERCURY.TRACER 58.01496 25.29168 2.294
## make.modelMITSUBISHI.GALANT 30.77546 9.31765 3.303
## make.modelMITSUBISHI.MIRAGE 47.69888 9.84559 4.845
## make.modelMITSUBISHI.MIRAGE DE 55.88755 33.16781 1.685
## make.modelNISSAN.ALTIMA 19.59745 8.82322 2.221
## make.modelNISSAN.MAXIMA -1.88096 9.46249 -0.199
## make.modelNISSAN.SENTRA 51.43826 8.84046 5.819
## make.modelNISSAN.STANZA 33.27050 16.67852 1.995
## make.modelOLDSMOBILE.88 34.34560 25.16320 1.365
## make.modelOLDSMOBILE.98 24.99953 25.46181 0.982
## make.modelOLDSMOBILE.ACHIEVA 57.47317 25.25125 2.276
## make.modelOLDSMOBILE.ALERO 23.57945 27.75001 0.850
## make.modelOLDSMOBILE.AURORA -9.84393 26.93639 -0.365
## make.modelOLDSMOBILE.CALAIS 32.93189 40.59901 0.811
## make.modelOLDSMOBILE.CIERA 64.95511 25.25214 2.572
## make.modelOLDSMOBILE.CUTLASS 24.49243 25.87185 0.947
## make.modelOLDSMOBILE.CUTLASS SUPREME 47.89598 25.12281 1.906
## make.modelOLDSMOBILE.INTRIGUE 7.93102 26.02295 0.305
## make.modelPLYMOUTH.ACCLAIM 61.30514 29.71046 2.063
## make.modelPLYMOUTH.BREEZE 28.98584 25.22934 1.149
## make.modelPLYMOUTH.NEON 46.50133 25.24045 1.842
## make.modelPONTIAC.BONNEVILLE 17.24033 25.08109 0.687
## make.modelPONTIAC.GRAND AM 31.50079 25.05340 1.257
## make.modelPONTIAC.GRAND PRIX 33.41614 25.16054 1.328
## make.modelPONTIAC.SUNBIRD 42.75998 31.17380 1.372
## make.modelPONTIAC.SUNFIRE 39.28965 25.35415 1.550
## make.modelSATURN.SL-1 16.85146 40.53595 0.416
## make.modelSUZUKI.ESTEEM 73.54765 10.81546 6.800
## make.modelTOYOTA.AVALON 9.93062 33.07160 0.300
## make.modelTOYOTA.CAMRY -4.07650 8.87621 -0.459
## make.modelTOYOTA.COROLLA 34.69596 8.97248 3.867
## make.modelTOYOTA.TERCEL 58.21554 9.46954 6.148
## s_total.cost 2.59981 0.07537 34.496
## s_mileage 0.13894 0.01456 9.541
## Pr(>|t|)
## (Intercept) < 2e-16 ***
## vehicle.typeFAMILY.LARGE 0.669164
## vehicle.typeFAMILY.MEDIUM 0.456436
## vehicle.typeFAMILY.SMALL 0.129043
## vehicle.typeLUXURY 0.605334
## domestic.importImport 0.759022
## vehicle.age < 2e-16 ***
## vehicle.age.group.L 0.255149
## vehicle.age.group.Q 0.877216
## vehicle.age.group.C 0.931422
## vehicle.age.group^4 0.231221
## color.setBLUE 0.086157 .
## color.setGOLD 0.880416
## color.setGREEN 0.087567 .
## color.setPURPLE 0.510948
## color.setRED 0.118410
## color.setSILVER 0.846977
## color.setWHITE 0.238561
## stateCA 1.38e-10 ***
## stateFL 0.082263 .
## stateGA < 2e-16 ***
## stateNM 4.43e-11 ***
## stateNV 0.117232
## stateTX 1.86e-14 ***
## stateVA 5.00e-07 ***
## make.modelBUICK.CENTURY 0.049165 *
## make.modelBUICK.LESABRE 0.207069
## make.modelBUICK.PARK AVENUE 0.384110
## make.modelBUICK.REGAL 0.151557
## make.modelBUICK.ROADMASTER 0.583550
## make.modelBUICK.SKYLARK 0.017783 *
## make.modelCADILLAC.BROUGHAM RWD 0.490137
## make.modelCADILLAC.CATERA 0.624427
## make.modelCADILLAC.DEVILLE 0.818510
## make.modelCADILLAC.FLEETWOOD RWD 0.428352
## make.modelCADILLAC.SEVILLE 0.315809
## make.modelCHEVROLET.CAPRICE 0.789266
## make.modelCHEVROLET.CAPRICE CLASSIC 0.948286
## make.modelCHEVROLET.CAVALIER 0.060494 .
## make.modelCHEVROLET.CELEBRITY 0.949127
## make.modelCHEVROLET.CORSICA 0.009575 **
## make.modelCHEVROLET.LUMINA 0.134101
## make.modelCHEVROLET.MALIBU 0.485716
## make.modelCHEVROLET.METRO 0.009052 **
## make.modelCHEVROLET.PRIZM 0.036057 *
## make.modelCHRYSLER.CIRRUS 0.847665
## make.modelCHRYSLER.CONCORDE 0.583566
## make.modelCHRYSLER.FIFTH AVENUE 0.606710
## make.modelCHRYSLER.LEBARON 0.402716
## make.modelCHRYSLER.LHS 0.899932
## make.modelCHRYSLER.NEW YORKER 0.717447
## make.modelDAEWOO.LANOS < 2e-16 ***
## make.modelDAEWOO.LEGANZA 0.113239
## make.modelDAEWOO.NUBIRA 3.71e-08 ***
## make.modelDAEWOO.NUBIRA SX 0.001770 **
## make.modelDODGE.DYNASTY 0.228641
## make.modelDODGE.INTREPID 0.515761
## make.modelDODGE.NEON 0.059762 .
## make.modelDODGE.SPIRIT 0.109593
## make.modelDODGE.STRATUS 0.529236
## make.modelEAGLE.VISION 0.409765
## make.modelEAGLE.VISION TSI 0.982932
## make.modelFORD.CONTOUR 0.202388
## make.modelFORD.CROWN VICTORIA 0.436230
## make.modelFORD.ESCORT 0.030979 *
## make.modelFORD.FOCUS 0.079183 .
## make.modelFORD.TAURUS 0.205026
## make.modelFORD.TEMPO 0.061886 .
## make.modelGEO.METRO 2.00e-13 ***
## make.modelGEO.PRIZM 3.58e-09 ***
## make.modelHONDA.ACCORD 0.539550
## make.modelHONDA.CIVIC 0.201043
## make.modelHYUNDAI.ACCENT 1.03e-11 ***
## make.modelHYUNDAI.ACCENT GL 0.401321
## make.modelHYUNDAI.ELANTRA 1.24e-09 ***
## make.modelHYUNDAI.SONATA 0.000105 ***
## make.modelINFINITI.G20 0.651192
## make.modelINFINITI.I30 0.349955
## make.modelINFINITI.J30 0.113918
## make.modelINFINITI.Q45 0.287465
## make.modelKIA.SEPHIA 3.50e-13 ***
## make.modelLEXUS.300 0.075286 .
## make.modelLINCOLN.TOWN CAR 0.998942
## make.modelMAZDA.626 0.025957 *
## make.modelMAZDA.MILLENIA 0.519823
## make.modelMAZDA.PROTEGE 1.01e-07 ***
## make.modelMERCURY.GRAND MARQUIS 0.702979
## make.modelMERCURY.MYSTIQUE 0.157562
## make.modelMERCURY.SABLE 0.175884
## make.modelMERCURY.TOPAZ 0.284551
## make.modelMERCURY.TRACER 0.021826 *
## make.modelMITSUBISHI.GALANT 0.000961 ***
## make.modelMITSUBISHI.MIRAGE 1.29e-06 ***
## make.modelMITSUBISHI.MIRAGE DE 0.092029 .
## make.modelNISSAN.ALTIMA 0.026371 *
## make.modelNISSAN.MAXIMA 0.842440
## make.modelNISSAN.SENTRA 6.17e-09 ***
## make.modelNISSAN.STANZA 0.046098 *
## make.modelOLDSMOBILE.88 0.172319
## make.modelOLDSMOBILE.98 0.326207
## make.modelOLDSMOBILE.ACHIEVA 0.022870 *
## make.modelOLDSMOBILE.ALERO 0.395513
## make.modelOLDSMOBILE.AURORA 0.714785
## make.modelOLDSMOBILE.CALAIS 0.417304
## make.modelOLDSMOBILE.CIERA 0.010122 *
## make.modelOLDSMOBILE.CUTLASS 0.343830
## make.modelOLDSMOBILE.CUTLASS SUPREME 0.056625 .
## make.modelOLDSMOBILE.INTRIGUE 0.760549
## make.modelPLYMOUTH.ACCLAIM 0.039106 *
## make.modelPLYMOUTH.BREEZE 0.250635
## make.modelPLYMOUTH.NEON 0.065464 .
## make.modelPONTIAC.BONNEVILLE 0.491861
## make.modelPONTIAC.GRAND AM 0.208666
## make.modelPONTIAC.GRAND PRIX 0.184178
## make.modelPONTIAC.SUNBIRD 0.170207
## make.modelPONTIAC.SUNFIRE 0.121270
## make.modelSATURN.SL-1 0.677629
## make.modelSUZUKI.ESTEEM 1.12e-11 ***
## make.modelTOYOTA.AVALON 0.763974
## make.modelTOYOTA.CAMRY 0.646060
## make.modelTOYOTA.COROLLA 0.000111 ***
## make.modelTOYOTA.TERCEL 8.25e-10 ***
## s_total.cost < 2e-16 ***
## s_mileage < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 31.88 on 7866 degrees of freedom
## Multiple R-squared: 0.2331, Adjusted R-squared: 0.2212
## F-statistic: 19.44 on 123 and 7866 DF, p-value: < 2.2e-16Plot the chart o show the variables contribute to R-squared and check VIF
options(repr.plot.width = 12, repr.plot.height = 4)
plot(calc.relimp(model2, rela=TRUE))
vif(model2)## GVIF Df GVIF^(1/(2*Df))
## vehicle.type 16.166273 4 1.416042
## domestic.import 934.341563 1 30.567001
## vehicle.age 16.552492 1 4.068475
## vehicle.age.group 18.940344 4 1.444353
## color.set 1.367567 7 1.022611
## state 2.092432 7 1.054153
## make.model 66992.476476 97 1.058952
## s_total.cost 2.016318 1 1.419971
## s_mileage 1.797834 1 1.340833Remove dosmetics.import due to instability of model and not significant in model2
model3 <- lm(lot.sale.days~. -total.cost -mileage -makex -domestic.import, data = train_set[,-1])
summary(model3)##
## Call:
## lm(formula = lot.sale.days ~ . - total.cost - mileage - makex -
## domestic.import, data = train_set[, -1])
##
## Residuals:
## Min 1Q Median 3Q Max
## -81.770 -22.291 -7.316 18.445 107.427
##
## Coefficients:
## Estimate Std. Error t value
## (Intercept) -254.22408 13.65684 -18.615
## vehicle.typeFAMILY.LARGE -0.91576 2.18733 -0.419
## vehicle.typeFAMILY.MEDIUM -1.40500 1.91290 -0.734
## vehicle.typeFAMILY.SMALL -2.46258 1.63682 -1.504
## vehicle.typeLUXURY -1.06424 2.09576 -0.508
## vehicle.age 7.09601 0.83303 8.518
## vehicle.age.group.L 3.56652 3.14377 1.134
## vehicle.age.group.Q 0.14576 0.92178 0.158
## vehicle.age.group.C 0.07534 0.86095 0.088
## vehicle.age.group^4 0.94482 0.78833 1.199
## color.setBLUE 2.82270 1.64344 1.718
## color.setGOLD -0.26623 1.77740 -0.150
## color.setGREEN 2.65174 1.55618 1.704
## color.setPURPLE 1.14910 1.74349 0.659
## color.setRED 2.91667 1.86574 1.563
## color.setSILVER 0.33618 1.73449 0.194
## color.setWHITE 1.85109 1.56836 1.180
## stateCA -8.89160 1.38389 -6.425
## stateFL -2.44330 1.40764 -1.736
## stateGA -18.46763 1.66594 -11.085
## stateNM -15.98518 2.42258 -6.598
## stateNV -3.76444 2.40564 -1.565
## stateTX -10.52001 1.37099 -7.673
## stateVA -8.78270 1.74665 -5.028
## make.modelBUICK.CENTURY 56.74336 9.28567 6.111
## make.modelBUICK.LESABRE 38.87659 9.00963 4.315
## make.modelBUICK.PARK AVENUE 29.32242 9.82990 2.983
## make.modelBUICK.REGAL 43.23274 9.08111 4.761
## make.modelBUICK.ROADMASTER 21.21979 10.27618 2.065
## make.modelBUICK.SKYLARK 66.94398 9.33779 7.169
## make.modelCADILLAC.BROUGHAM RWD -20.75796 33.02432 -0.629
## make.modelCADILLAC.CATERA -8.00527 20.34756 -0.393
## make.modelCADILLAC.DEVILLE 1.39340 9.43325 0.148
## make.modelCADILLAC.FLEETWOOD RWD -15.61569 16.72426 -0.934
## make.modelCADILLAC.SEVILLE -19.82331 13.24581 -1.497
## make.modelCHEVROLET.CAPRICE 16.19467 24.16979 0.670
## make.modelCHEVROLET.CAPRICE CLASSIC 8.86960 10.99546 0.807
## make.modelCHEVROLET.CAVALIER 54.30356 9.11759 5.956
## make.modelCHEVROLET.CELEBRITY 9.79124 33.20101 0.295
## make.modelCHEVROLET.CORSICA 72.80382 9.60910 7.577
## make.modelCHEVROLET.LUMINA 44.72801 8.92792 5.010
## make.modelCHEVROLET.MALIBU 24.68334 9.17245 2.691
## make.modelCHEVROLET.METRO 74.63789 10.93161 6.828
## make.modelCHEVROLET.PRIZM 57.55195 10.60297 5.428
## make.modelCHRYSLER.CIRRUS 12.06435 9.80497 1.230
## make.modelCHRYSLER.CONCORDE 21.04423 9.49236 2.217
## make.modelCHRYSLER.FIFTH AVENUE -13.67204 33.05804 -0.414
## make.modelCHRYSLER.LEBARON 41.15645 33.14232 1.242
## make.modelCHRYSLER.LHS 10.39424 10.02659 1.037
## make.modelCHRYSLER.NEW YORKER 16.82935 12.67800 1.327
## make.modelDAEWOO.LANOS 116.44448 11.44321 10.176
## make.modelDAEWOO.LEGANZA 32.52081 20.54880 1.583
## make.modelDAEWOO.NUBIRA 66.29259 12.03838 5.507
## make.modelDAEWOO.NUBIRA SX 103.78996 33.18758 3.127
## make.modelDODGE.DYNASTY 56.14530 33.15726 1.693
## make.modelDODGE.INTREPID 23.49079 9.03142 2.601
## make.modelDODGE.NEON 54.57092 9.23937 5.906
## make.modelDODGE.SPIRIT 61.25796 24.30909 2.520
## make.modelDODGE.STRATUS 23.00866 9.28431 2.478
## make.modelEAGLE.VISION 28.60716 11.25969 2.541
## make.modelEAGLE.VISION TSI 6.32166 33.04787 0.191
## make.modelFORD.CONTOUR 39.15291 9.07091 4.316
## make.modelFORD.CROWN VICTORIA 26.84623 9.34998 2.871
## make.modelFORD.ESCORT 61.45993 9.14981 6.717
## make.modelFORD.FOCUS 78.30964 33.14889 2.362
## make.modelFORD.TAURUS 38.91542 8.93243 4.357
## make.modelFORD.TEMPO 56.92209 12.56679 4.530
## make.modelGEO.METRO 78.84299 10.71006 7.362
## make.modelGEO.PRIZM 54.28202 9.18679 5.909
## make.modelHONDA.ACCORD -5.68468 9.26779 -0.613
## make.modelHONDA.CIVIC 11.78191 9.21786 1.278
## make.modelHYUNDAI.ACCENT 68.15346 10.00623 6.811
## make.modelHYUNDAI.ACCENT GL 27.77713 33.12687 0.839
## make.modelHYUNDAI.ELANTRA 59.76913 9.82934 6.081
## make.modelHYUNDAI.SONATA 40.87857 10.53664 3.880
## make.modelINFINITI.G20 4.70565 10.40549 0.452
## make.modelINFINITI.I30 22.58817 24.16820 0.935
## make.modelINFINITI.J30 -23.39682 14.79865 -1.581
## make.modelINFINITI.Q45 -21.59651 20.30227 -1.064
## make.modelKIA.SEPHIA 68.17556 9.35887 7.285
## make.modelLEXUS.300 -58.82273 33.06623 -1.779
## make.modelLINCOLN.TOWN CAR 7.16083 9.98632 0.717
## make.modelMAZDA.626 20.20184 9.07294 2.227
## make.modelMAZDA.MILLENIA -9.55387 14.84136 -0.644
## make.modelMAZDA.PROTEGE 48.54135 9.10867 5.329
## make.modelMERCURY.GRAND MARQUIS 16.81734 9.39234 1.791
## make.modelMERCURY.MYSTIQUE 42.99876 9.75173 4.409
## make.modelMERCURY.SABLE 41.31263 9.38901 4.400
## make.modelMERCURY.TOPAZ 40.47134 20.47739 1.976
## make.modelMERCURY.TRACER 65.20167 9.53188 6.840
## make.modelMITSUBISHI.GALANT 30.76763 9.31708 3.302
## make.modelMITSUBISHI.MIRAGE 47.68440 9.84491 4.844
## make.modelMITSUBISHI.MIRAGE DE 55.84387 33.16559 1.684
## make.modelNISSAN.ALTIMA 19.59094 8.82268 2.221
## make.modelNISSAN.MAXIMA -1.87948 9.46194 -0.199
## make.modelNISSAN.SENTRA 51.43150 8.83992 5.818
## make.modelNISSAN.STANZA 33.26559 16.67755 1.995
## make.modelOLDSMOBILE.88 41.53502 9.16306 4.533
## make.modelOLDSMOBILE.98 32.19253 9.92617 3.243
## make.modelOLDSMOBILE.ACHIEVA 64.65111 9.49478 6.809
## make.modelOLDSMOBILE.ALERO 30.74356 14.98935 2.051
## make.modelOLDSMOBILE.AURORA -2.65718 13.29439 -0.200
## make.modelOLDSMOBILE.CALAIS 40.11987 33.15354 1.210
## make.modelOLDSMOBILE.CIERA 72.13951 9.44515 7.638
## make.modelOLDSMOBILE.CUTLASS 31.67037 11.03964 2.869
## make.modelOLDSMOBILE.CUTLASS SUPREME 55.08547 9.05096 6.086
## make.modelOLDSMOBILE.INTRIGUE 15.10210 11.43506 1.321
## make.modelPLYMOUTH.ACCLAIM 68.49328 18.26616 3.750
## make.modelPLYMOUTH.BREEZE 36.15925 9.47290 3.817
## make.modelPLYMOUTH.NEON 53.68722 9.40180 5.710
## make.modelPONTIAC.BONNEVILLE 24.43101 8.92436 2.738
## make.modelPONTIAC.GRAND AM 38.68014 8.94339 4.325
## make.modelPONTIAC.GRAND PRIX 40.60111 9.19272 4.417
## make.modelPONTIAC.SUNBIRD 49.96526 20.49675 2.438
## make.modelPONTIAC.SUNFIRE 46.47551 9.70312 4.790
## make.modelSATURN.SL-1 24.02469 33.11028 0.726
## make.modelSUZUKI.ESTEEM 73.53015 10.81468 6.799
## make.modelTOYOTA.AVALON 9.94816 33.06965 0.301
## make.modelTOYOTA.CAMRY -4.07577 8.87570 -0.459
## make.modelTOYOTA.COROLLA 34.68763 8.97192 3.866
## make.modelTOYOTA.TERCEL 58.21898 9.46899 6.148
## s_total.cost 2.59980 0.07536 34.498
## s_mileage 0.13884 0.01456 9.537
## Pr(>|t|)
## (Intercept) < 2e-16 ***
## vehicle.typeFAMILY.LARGE 0.675474
## vehicle.typeFAMILY.MEDIUM 0.462676
## vehicle.typeFAMILY.SMALL 0.132494
## vehicle.typeLUXURY 0.611605
## vehicle.age < 2e-16 ***
## vehicle.age.group.L 0.256631
## vehicle.age.group.Q 0.874364
## vehicle.age.group.C 0.930271
## vehicle.age.group^4 0.230757
## color.setBLUE 0.085916 .
## color.setGOLD 0.880935
## color.setGREEN 0.088419 .
## color.setPURPLE 0.509861
## color.setRED 0.118028
## color.setSILVER 0.846319
## color.setWHITE 0.237928
## stateCA 1.39e-10 ***
## stateFL 0.082650 .
## stateGA < 2e-16 ***
## stateNM 4.42e-11 ***
## stateNV 0.117661
## stateTX 1.88e-14 ***
## stateVA 5.06e-07 ***
## make.modelBUICK.CENTURY 1.04e-09 ***
## make.modelBUICK.LESABRE 1.62e-05 ***
## make.modelBUICK.PARK AVENUE 0.002863 **
## make.modelBUICK.REGAL 1.96e-06 ***
## make.modelBUICK.ROADMASTER 0.038960 *
## make.modelBUICK.SKYLARK 8.23e-13 ***
## make.modelCADILLAC.BROUGHAM RWD 0.529652
## make.modelCADILLAC.CATERA 0.694015
## make.modelCADILLAC.DEVILLE 0.882574
## make.modelCADILLAC.FLEETWOOD RWD 0.350480
## make.modelCADILLAC.SEVILLE 0.134545
## make.modelCHEVROLET.CAPRICE 0.502854
## make.modelCHEVROLET.CAPRICE CLASSIC 0.419887
## make.modelCHEVROLET.CAVALIER 2.70e-09 ***
## make.modelCHEVROLET.CELEBRITY 0.768072
## make.modelCHEVROLET.CORSICA 3.95e-14 ***
## make.modelCHEVROLET.LUMINA 5.56e-07 ***
## make.modelCHEVROLET.MALIBU 0.007138 **
## make.modelCHEVROLET.METRO 9.27e-12 ***
## make.modelCHEVROLET.PRIZM 5.87e-08 ***
## make.modelCHRYSLER.CIRRUS 0.218572
## make.modelCHRYSLER.CONCORDE 0.026654 *
## make.modelCHRYSLER.FIFTH AVENUE 0.679195
## make.modelCHRYSLER.LEBARON 0.214344
## make.modelCHRYSLER.LHS 0.299923
## make.modelCHRYSLER.NEW YORKER 0.184400
## make.modelDAEWOO.LANOS < 2e-16 ***
## make.modelDAEWOO.LEGANZA 0.113550
## make.modelDAEWOO.NUBIRA 3.77e-08 ***
## make.modelDAEWOO.NUBIRA SX 0.001770 **
## make.modelDODGE.DYNASTY 0.090437 .
## make.modelDODGE.INTREPID 0.009312 **
## make.modelDODGE.NEON 3.64e-09 ***
## make.modelDODGE.SPIRIT 0.011756 *
## make.modelDODGE.STRATUS 0.013224 *
## make.modelEAGLE.VISION 0.011083 *
## make.modelEAGLE.VISION TSI 0.848305
## make.modelFORD.CONTOUR 1.61e-05 ***
## make.modelFORD.CROWN VICTORIA 0.004099 **
## make.modelFORD.ESCORT 1.98e-11 ***
## make.modelFORD.FOCUS 0.018183 *
## make.modelFORD.TAURUS 1.34e-05 ***
## make.modelFORD.TEMPO 6.00e-06 ***
## make.modelGEO.METRO 2.00e-13 ***
## make.modelGEO.PRIZM 3.59e-09 ***
## make.modelHONDA.ACCORD 0.539642
## make.modelHONDA.CIVIC 0.201230
## make.modelHYUNDAI.ACCENT 1.04e-11 ***
## make.modelHYUNDAI.ACCENT GL 0.401771
## make.modelHYUNDAI.ELANTRA 1.25e-09 ***
## make.modelHYUNDAI.SONATA 0.000105 ***
## make.modelINFINITI.G20 0.651117
## make.modelINFINITI.I30 0.350011
## make.modelINFINITI.J30 0.113916
## make.modelINFINITI.Q45 0.287475
## make.modelKIA.SEPHIA 3.54e-13 ***
## make.modelLEXUS.300 0.075289 .
## make.modelLINCOLN.TOWN CAR 0.473356
## make.modelMAZDA.626 0.026002 *
## make.modelMAZDA.MILLENIA 0.519768
## make.modelMAZDA.PROTEGE 1.01e-07 ***
## make.modelMERCURY.GRAND MARQUIS 0.073406 .
## make.modelMERCURY.MYSTIQUE 1.05e-05 ***
## make.modelMERCURY.SABLE 1.10e-05 ***
## make.modelMERCURY.TOPAZ 0.048145 *
## make.modelMERCURY.TRACER 8.49e-12 ***
## make.modelMITSUBISHI.GALANT 0.000963 ***
## make.modelMITSUBISHI.MIRAGE 1.30e-06 ***
## make.modelMITSUBISHI.MIRAGE DE 0.092262 .
## make.modelNISSAN.ALTIMA 0.026412 *
## make.modelNISSAN.MAXIMA 0.842553
## make.modelNISSAN.SENTRA 6.19e-09 ***
## make.modelNISSAN.STANZA 0.046117 *
## make.modelOLDSMOBILE.88 5.90e-06 ***
## make.modelOLDSMOBILE.98 0.001187 **
## make.modelOLDSMOBILE.ACHIEVA 1.05e-11 ***
## make.modelOLDSMOBILE.ALERO 0.040297 *
## make.modelOLDSMOBILE.AURORA 0.841586
## make.modelOLDSMOBILE.CALAIS 0.226268
## make.modelOLDSMOBILE.CIERA 2.47e-14 ***
## make.modelOLDSMOBILE.CUTLASS 0.004131 **
## make.modelOLDSMOBILE.CUTLASS SUPREME 1.21e-09 ***
## make.modelOLDSMOBILE.INTRIGUE 0.186645
## make.modelPLYMOUTH.ACCLAIM 0.000178 ***
## make.modelPLYMOUTH.BREEZE 0.000136 ***
## make.modelPLYMOUTH.NEON 1.17e-08 ***
## make.modelPONTIAC.BONNEVILLE 0.006203 **
## make.modelPONTIAC.GRAND AM 1.54e-05 ***
## make.modelPONTIAC.GRAND PRIX 1.02e-05 ***
## make.modelPONTIAC.SUNBIRD 0.014802 *
## make.modelPONTIAC.SUNFIRE 1.70e-06 ***
## make.modelSATURN.SL-1 0.468108
## make.modelSUZUKI.ESTEEM 1.13e-11 ***
## make.modelTOYOTA.AVALON 0.763556
## make.modelTOYOTA.CAMRY 0.646100
## make.modelTOYOTA.COROLLA 0.000111 ***
## make.modelTOYOTA.TERCEL 8.21e-10 ***
## s_total.cost < 2e-16 ***
## s_mileage < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 31.88 on 7867 degrees of freedom
## Multiple R-squared: 0.2331, Adjusted R-squared: 0.2212
## F-statistic: 19.6 on 122 and 7867 DF, p-value: < 2.2e-16options(repr.plot.width = 12, repr.plot.height = 4)
plot(calc.relimp(model3, rela=TRUE))
vif(model3)## GVIF Df GVIF^(1/(2*Df))
## vehicle.type 16.124182 4 1.415581
## vehicle.age 16.551466 1 4.068349
## vehicle.age.group 18.932301 4 1.444277
## color.set 1.365898 7 1.022522
## state 2.087753 7 1.053985
## make.model 106.168719 97 1.024338
## s_total.cost 2.016317 1 1.419971
## s_mileage 1.796973 1 1.340512Remove vehicle.age.group due to instability of model and not significant in model3
we can use vehicle.age to predict instead of vehicle.age.group
model4 <- lm(lot.sale.days~. -total.cost -mileage -makex - domestic.import -vehicle.age.group, data = train_set[,-1])
summary(model4)##
## Call:
## lm(formula = lot.sale.days ~ . - total.cost - mileage - makex -
## domestic.import - vehicle.age.group, data = train_set[, -1])
##
## Residuals:
## Min 1Q Median 3Q Max
## -83.429 -22.183 -7.375 18.475 107.382
##
## Coefficients:
## Estimate Std. Error t value
## (Intercept) -259.09270 12.90251 -20.081
## vehicle.typeFAMILY.LARGE -0.98309 2.18653 -0.450
## vehicle.typeFAMILY.MEDIUM -1.41826 1.91216 -0.742
## vehicle.typeFAMILY.SMALL -2.48307 1.63465 -1.519
## vehicle.typeLUXURY -1.11762 2.09405 -0.534
## vehicle.age 7.95748 0.35586 22.361
## color.setBLUE 2.82609 1.64257 1.721
## color.setGOLD -0.29327 1.77576 -0.165
## color.setGREEN 2.66560 1.55555 1.714
## color.setPURPLE 1.12226 1.74313 0.644
## color.setRED 2.89880 1.86520 1.554
## color.setSILVER 0.29281 1.73361 0.169
## color.setWHITE 1.83641 1.56806 1.171
## stateCA -8.89043 1.38367 -6.425
## stateFL -2.41875 1.40593 -1.720
## stateGA -18.48675 1.66553 -11.100
## stateNM -15.98479 2.42180 -6.600
## stateNV -3.67194 2.40452 -1.527
## stateTX -10.54985 1.37066 -7.697
## stateVA -8.81265 1.74546 -5.049
## make.modelBUICK.CENTURY 57.24099 9.26169 6.180
## make.modelBUICK.LESABRE 39.48134 8.98681 4.393
## make.modelBUICK.PARK AVENUE 29.96032 9.80886 3.054
## make.modelBUICK.REGAL 43.92851 9.05111 4.853
## make.modelBUICK.ROADMASTER 21.51532 10.27105 2.095
## make.modelBUICK.SKYLARK 67.46078 9.31387 7.243
## make.modelCADILLAC.BROUGHAM RWD -21.16909 33.01958 -0.641
## make.modelCADILLAC.CATERA -7.27390 20.33120 -0.358
## make.modelCADILLAC.DEVILLE 1.66857 9.42857 0.177
## make.modelCADILLAC.FLEETWOOD RWD -15.14099 16.71743 -0.906
## make.modelCADILLAC.SEVILLE -19.95501 13.24437 -1.507
## make.modelCHEVROLET.CAPRICE 15.84918 24.16590 0.656
## make.modelCHEVROLET.CAPRICE CLASSIC 9.62287 10.97129 0.877
## make.modelCHEVROLET.CAVALIER 54.67335 9.10125 6.007
## make.modelCHEVROLET.CELEBRITY 6.91332 33.10530 0.209
## make.modelCHEVROLET.CORSICA 73.48817 9.57728 7.673
## make.modelCHEVROLET.LUMINA 45.28209 8.89970 5.088
## make.modelCHEVROLET.MALIBU 24.85143 9.16369 2.712
## make.modelCHEVROLET.METRO 74.79884 10.92316 6.848
## make.modelCHEVROLET.PRIZM 57.66748 10.59526 5.443
## make.modelCHRYSLER.CIRRUS 12.79621 9.76375 1.311
## make.modelCHRYSLER.CONCORDE 21.79962 9.45618 2.305
## make.modelCHRYSLER.FIFTH AVENUE -14.05622 33.05380 -0.425
## make.modelCHRYSLER.LEBARON 39.96403 33.12428 1.206
## make.modelCHRYSLER.LHS 11.25455 9.98404 1.127
## make.modelCHRYSLER.NEW YORKER 18.01797 12.62951 1.427
## make.modelDAEWOO.LANOS 118.24968 11.32170 10.445
## make.modelDAEWOO.LEGANZA 33.82623 20.50979 1.649
## make.modelDAEWOO.NUBIRA 68.09063 11.92318 5.711
## make.modelDAEWOO.NUBIRA SX 103.94779 33.18016 3.133
## make.modelDODGE.DYNASTY 54.94073 33.13812 1.658
## make.modelDODGE.INTREPID 24.13372 9.00000 2.682
## make.modelDODGE.NEON 55.02975 9.21604 5.971
## make.modelDODGE.SPIRIT 62.12430 24.28027 2.559
## make.modelDODGE.STRATUS 23.49387 9.26705 2.535
## make.modelEAGLE.VISION 29.05400 11.23461 2.586
## make.modelEAGLE.VISION TSI 6.67813 33.02394 0.202
## make.modelFORD.CONTOUR 39.57237 9.05079 4.372
## make.modelFORD.CROWN VICTORIA 27.36661 9.33370 2.932
## make.modelFORD.ESCORT 61.71898 9.13089 6.759
## make.modelFORD.FOCUS 79.37666 33.12917 2.396
## make.modelFORD.TAURUS 39.37486 8.90968 4.419
## make.modelFORD.TEMPO 57.17835 12.56239 4.552
## make.modelGEO.METRO 79.00495 10.67915 7.398
## make.modelGEO.PRIZM 54.96210 9.14977 6.007
## make.modelHONDA.ACCORD -5.33594 9.25460 -0.577
## make.modelHONDA.CIVIC 12.36192 9.19712 1.344
## make.modelHYUNDAI.ACCENT 68.52519 9.99543 6.856
## make.modelHYUNDAI.ACCENT GL 27.61395 33.11310 0.834
## make.modelHYUNDAI.ELANTRA 60.06279 9.81988 6.116
## make.modelHYUNDAI.SONATA 41.08860 10.52819 3.903
## make.modelINFINITI.G20 5.13202 10.39631 0.494
## make.modelINFINITI.I30 23.31638 24.14962 0.965
## make.modelINFINITI.J30 -22.94583 14.79002 -1.551
## make.modelINFINITI.Q45 -21.41123 20.29971 -1.055
## make.modelKIA.SEPHIA 68.58817 9.34278 7.341
## make.modelLEXUS.300 -59.23068 33.06184 -1.792
## make.modelLINCOLN.TOWN CAR 7.69748 9.97114 0.772
## make.modelMAZDA.626 20.86899 9.04014 2.308
## make.modelMAZDA.MILLENIA -8.65182 14.80559 -0.584
## make.modelMAZDA.PROTEGE 49.01581 9.07794 5.399
## make.modelMERCURY.GRAND MARQUIS 17.32908 9.37751 1.848
## make.modelMERCURY.MYSTIQUE 43.50823 9.72651 4.473
## make.modelMERCURY.SABLE 41.84737 9.35918 4.471
## make.modelMERCURY.TOPAZ 39.86052 20.46973 1.947
## make.modelMERCURY.TRACER 65.51966 9.51297 6.887
## make.modelMITSUBISHI.GALANT 31.44017 9.27861 3.388
## make.modelMITSUBISHI.MIRAGE 47.96886 9.83159 4.879
## make.modelMITSUBISHI.MIRAGE DE 56.85418 33.14601 1.715
## make.modelNISSAN.ALTIMA 20.27796 8.79008 2.307
## make.modelNISSAN.MAXIMA -1.23579 9.43882 -0.131
## make.modelNISSAN.SENTRA 52.08465 8.80927 5.912
## make.modelNISSAN.STANZA 32.88696 16.67328 1.972
## make.modelOLDSMOBILE.88 42.08622 9.13818 4.606
## make.modelOLDSMOBILE.98 32.48395 9.91929 3.275
## make.modelOLDSMOBILE.ACHIEVA 65.00177 9.47219 6.862
## make.modelOLDSMOBILE.ALERO 31.43186 14.96618 2.100
## make.modelOLDSMOBILE.AURORA -2.10475 13.25805 -0.159
## make.modelOLDSMOBILE.CALAIS 38.96295 33.13626 1.176
## make.modelOLDSMOBILE.CIERA 72.72743 9.41661 7.723
## make.modelOLDSMOBILE.CUTLASS 31.86043 11.02147 2.891
## make.modelOLDSMOBILE.CUTLASS SUPREME 55.78898 9.01799 6.186
## make.modelOLDSMOBILE.INTRIGUE 15.12531 11.42887 1.323
## make.modelPLYMOUTH.ACCLAIM 68.75623 18.25699 3.766
## make.modelPLYMOUTH.BREEZE 36.47704 9.45894 3.856
## make.modelPLYMOUTH.NEON 54.11076 9.38117 5.768
## make.modelPONTIAC.BONNEVILLE 25.00585 8.89571 2.811
## make.modelPONTIAC.GRAND AM 39.26018 8.91532 4.404
## make.modelPONTIAC.GRAND PRIX 41.20894 9.16597 4.496
## make.modelPONTIAC.SUNBIRD 49.92275 20.49461 2.436
## make.modelPONTIAC.SUNFIRE 46.73775 9.68474 4.826
## make.modelSATURN.SL-1 25.26121 33.09066 0.763
## make.modelSUZUKI.ESTEEM 73.83097 10.80033 6.836
## make.modelTOYOTA.AVALON 11.33038 33.04537 0.343
## make.modelTOYOTA.CAMRY -3.57129 8.86110 -0.403
## make.modelTOYOTA.COROLLA 35.33670 8.94777 3.949
## make.modelTOYOTA.TERCEL 58.80966 9.43072 6.236
## s_total.cost 2.60189 0.07528 34.563
## s_mileage 0.13841 0.01451 9.542
## Pr(>|t|)
## (Intercept) < 2e-16 ***
## vehicle.typeFAMILY.LARGE 0.653003
## vehicle.typeFAMILY.MEDIUM 0.458289
## vehicle.typeFAMILY.SMALL 0.128798
## vehicle.typeLUXURY 0.593556
## vehicle.age < 2e-16 ***
## color.setBLUE 0.085376 .
## color.setGOLD 0.868830
## color.setGREEN 0.086640 .
## color.setPURPLE 0.519712
## color.setRED 0.120189
## color.setSILVER 0.865878
## color.setWHITE 0.241582
## stateCA 1.39e-10 ***
## stateFL 0.085400 .
## stateGA < 2e-16 ***
## stateNM 4.37e-11 ***
## stateNV 0.126778
## stateTX 1.56e-14 ***
## stateVA 4.54e-07 ***
## make.modelBUICK.CENTURY 6.71e-10 ***
## make.modelBUICK.LESABRE 1.13e-05 ***
## make.modelBUICK.PARK AVENUE 0.002263 **
## make.modelBUICK.REGAL 1.24e-06 ***
## make.modelBUICK.ROADMASTER 0.036225 *
## make.modelBUICK.SKYLARK 4.80e-13 ***
## make.modelCADILLAC.BROUGHAM RWD 0.521472
## make.modelCADILLAC.CATERA 0.720525
## make.modelCADILLAC.DEVILLE 0.859537
## make.modelCADILLAC.FLEETWOOD RWD 0.365122
## make.modelCADILLAC.SEVILLE 0.131933
## make.modelCHEVROLET.CAPRICE 0.511941
## make.modelCHEVROLET.CAPRICE CLASSIC 0.380462
## make.modelCHEVROLET.CAVALIER 1.97e-09 ***
## make.modelCHEVROLET.CELEBRITY 0.834588
## make.modelCHEVROLET.CORSICA 1.88e-14 ***
## make.modelCHEVROLET.LUMINA 3.70e-07 ***
## make.modelCHEVROLET.MALIBU 0.006703 **
## make.modelCHEVROLET.METRO 8.07e-12 ***
## make.modelCHEVROLET.PRIZM 5.40e-08 ***
## make.modelCHRYSLER.CIRRUS 0.190037
## make.modelCHRYSLER.CONCORDE 0.021174 *
## make.modelCHRYSLER.FIFTH AVENUE 0.670664
## make.modelCHRYSLER.LEBARON 0.227666
## make.modelCHRYSLER.LHS 0.259670
## make.modelCHRYSLER.NEW YORKER 0.153719
## make.modelDAEWOO.LANOS < 2e-16 ***
## make.modelDAEWOO.LEGANZA 0.099132 .
## make.modelDAEWOO.NUBIRA 1.17e-08 ***
## make.modelDAEWOO.NUBIRA SX 0.001738 **
## make.modelDODGE.DYNASTY 0.097371 .
## make.modelDODGE.INTREPID 0.007344 **
## make.modelDODGE.NEON 2.46e-09 ***
## make.modelDODGE.SPIRIT 0.010527 *
## make.modelDODGE.STRATUS 0.011257 *
## make.modelEAGLE.VISION 0.009724 **
## make.modelEAGLE.VISION TSI 0.839749
## make.modelFORD.CONTOUR 1.25e-05 ***
## make.modelFORD.CROWN VICTORIA 0.003377 **
## make.modelFORD.ESCORT 1.48e-11 ***
## make.modelFORD.FOCUS 0.016599 *
## make.modelFORD.TAURUS 1.00e-05 ***
## make.modelFORD.TEMPO 5.41e-06 ***
## make.modelGEO.METRO 1.52e-13 ***
## make.modelGEO.PRIZM 1.97e-09 ***
## make.modelHONDA.ACCORD 0.564246
## make.modelHONDA.CIVIC 0.178952
## make.modelHYUNDAI.ACCENT 7.64e-12 ***
## make.modelHYUNDAI.ACCENT GL 0.404347
## make.modelHYUNDAI.ELANTRA 1.00e-09 ***
## make.modelHYUNDAI.SONATA 9.59e-05 ***
## make.modelINFINITI.G20 0.621575
## make.modelINFINITI.I30 0.334326
## make.modelINFINITI.J30 0.120837
## make.modelINFINITI.Q45 0.291570
## make.modelKIA.SEPHIA 2.33e-13 ***
## make.modelLEXUS.300 0.073250 .
## make.modelLINCOLN.TOWN CAR 0.440152
## make.modelMAZDA.626 0.020998 *
## make.modelMAZDA.MILLENIA 0.558994
## make.modelMAZDA.PROTEGE 6.88e-08 ***
## make.modelMERCURY.GRAND MARQUIS 0.064649 .
## make.modelMERCURY.MYSTIQUE 7.82e-06 ***
## make.modelMERCURY.SABLE 7.88e-06 ***
## make.modelMERCURY.TOPAZ 0.051535 .
## make.modelMERCURY.TRACER 6.12e-12 ***
## make.modelMITSUBISHI.GALANT 0.000706 ***
## make.modelMITSUBISHI.MIRAGE 1.09e-06 ***
## make.modelMITSUBISHI.MIRAGE DE 0.086336 .
## make.modelNISSAN.ALTIMA 0.021085 *
## make.modelNISSAN.MAXIMA 0.895837
## make.modelNISSAN.SENTRA 3.51e-09 ***
## make.modelNISSAN.STANZA 0.048595 *
## make.modelOLDSMOBILE.88 4.18e-06 ***
## make.modelOLDSMOBILE.98 0.001062 **
## make.modelOLDSMOBILE.ACHIEVA 7.29e-12 ***
## make.modelOLDSMOBILE.ALERO 0.035744 *
## make.modelOLDSMOBILE.AURORA 0.873868
## make.modelOLDSMOBILE.CALAIS 0.239694
## make.modelOLDSMOBILE.CIERA 1.27e-14 ***
## make.modelOLDSMOBILE.CUTLASS 0.003854 **
## make.modelOLDSMOBILE.CUTLASS SUPREME 6.46e-10 ***
## make.modelOLDSMOBILE.INTRIGUE 0.185731
## make.modelPLYMOUTH.ACCLAIM 0.000167 ***
## make.modelPLYMOUTH.BREEZE 0.000116 ***
## make.modelPLYMOUTH.NEON 8.32e-09 ***
## make.modelPONTIAC.BONNEVILLE 0.004951 **
## make.modelPONTIAC.GRAND AM 1.08e-05 ***
## make.modelPONTIAC.GRAND PRIX 7.03e-06 ***
## make.modelPONTIAC.SUNBIRD 0.014877 *
## make.modelPONTIAC.SUNFIRE 1.42e-06 ***
## make.modelSATURN.SL-1 0.445251
## make.modelSUZUKI.ESTEEM 8.75e-12 ***
## make.modelTOYOTA.AVALON 0.731703
## make.modelTOYOTA.CAMRY 0.686937
## make.modelTOYOTA.COROLLA 7.91e-05 ***
## make.modelTOYOTA.TERCEL 4.72e-10 ***
## s_total.cost < 2e-16 ***
## s_mileage < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 31.87 on 7871 degrees of freedom
## Multiple R-squared: 0.2329, Adjusted R-squared: 0.2214
## F-statistic: 20.25 on 118 and 7871 DF, p-value: < 2.2e-16options(repr.plot.width = 12, repr.plot.height = 4)
plot(calc.relimp(model4, rela=TRUE))
vif(model4)## GVIF Df GVIF^(1/(2*Df))
## vehicle.type 16.040930 4 1.414665
## vehicle.age 3.021026 1 1.738110
## color.set 1.356699 7 1.022029
## state 2.067070 7 1.053235
## make.model 76.495680 97 1.022609
## s_total.cost 2.012276 1 1.418547
## s_mileage 1.784445 1 1.335831Remove the vehicle.type since it is not significant in model4 and not model is still not stable.
model5 <- lm(lot.sale.days~. -total.cost -mileage -makex - domestic.import -vehicle.age.group -vehicle.type, data = train_set[,-1])
summary(model5)##
## Call:
## lm(formula = lot.sale.days ~ . - total.cost - mileage - makex -
## domestic.import - vehicle.age.group - vehicle.type, data = train_set[,
## -1])
##
## Residuals:
## Min 1Q Median 3Q Max
## -84.540 -22.232 -7.353 18.453 107.513
##
## Coefficients:
## Estimate Std. Error t value
## (Intercept) -259.87286 12.66210 -20.524
## vehicle.age 7.96084 0.35553 22.391
## color.setBLUE 2.81291 1.64191 1.713
## color.setGOLD -0.31008 1.77538 -0.175
## color.setGREEN 2.64016 1.55473 1.698
## color.setPURPLE 1.08105 1.74192 0.621
## color.setRED 2.95798 1.86434 1.587
## color.setSILVER 0.30430 1.73328 0.176
## color.setWHITE 1.79756 1.56735 1.147
## stateCA -8.92708 1.28838 -6.929
## stateFL -2.21889 1.33553 -1.661
## stateGA -18.16992 1.57140 -11.563
## stateNM -15.97624 2.38129 -6.709
## stateNV -3.24265 2.34565 -1.382
## stateTX -10.29694 1.26775 -8.122
## stateVA -8.58277 1.69828 -5.054
## make.modelBUICK.CENTURY 57.07858 9.14640 6.241
## make.modelBUICK.LESABRE 39.54507 8.83452 4.476
## make.modelBUICK.PARK AVENUE 29.89182 9.75756 3.063
## make.modelBUICK.REGAL 43.71328 8.92434 4.898
## make.modelBUICK.ROADMASTER 21.56567 10.19171 2.116
## make.modelBUICK.SKYLARK 67.15064 9.20923 7.292
## make.modelCADILLAC.BROUGHAM RWD -21.21085 33.01619 -0.642
## make.modelCADILLAC.CATERA -7.29706 20.32698 -0.359
## make.modelCADILLAC.DEVILLE 1.69550 9.38830 0.181
## make.modelCADILLAC.FLEETWOOD RWD -15.00588 16.64308 -0.902
## make.modelCADILLAC.SEVILLE -19.98263 13.21269 -1.512
## make.modelCHEVROLET.CAPRICE 15.79995 24.12233 0.655
## make.modelCHEVROLET.CAPRICE CLASSIC 9.72515 10.83522 0.898
## make.modelCHEVROLET.CAVALIER 54.12849 8.98582 6.024
## make.modelCHEVROLET.CELEBRITY 6.47553 33.06731 0.196
## make.modelCHEVROLET.CORSICA 73.35606 9.47520 7.742
## make.modelCHEVROLET.LUMINA 45.20971 8.76824 5.156
## make.modelCHEVROLET.MALIBU 24.53514 9.03430 2.716
## make.modelCHEVROLET.METRO 75.13371 10.77741 6.971
## make.modelCHEVROLET.PRIZM 56.97874 10.48036 5.437
## make.modelCHRYSLER.CIRRUS 12.57452 9.65006 1.303
## make.modelCHRYSLER.CONCORDE 21.85084 9.34628 2.338
## make.modelCHRYSLER.FIFTH AVENUE -14.23605 33.02879 -0.431
## make.modelCHRYSLER.LEBARON 39.51613 33.08563 1.194
## make.modelCHRYSLER.LHS 11.31787 9.94485 1.138
## make.modelCHRYSLER.NEW YORKER 17.93977 12.57780 1.426
## make.modelDAEWOO.LANOS 117.92806 11.22885 10.502
## make.modelDAEWOO.LEGANZA 33.90027 20.45343 1.657
## make.modelDAEWOO.NUBIRA 67.66331 11.83209 5.719
## make.modelDAEWOO.NUBIRA SX 103.44228 33.13614 3.122
## make.modelDODGE.DYNASTY 54.39946 33.09464 1.644
## make.modelDODGE.INTREPID 24.14473 8.94161 2.700
## make.modelDODGE.NEON 55.19465 9.06580 6.088
## make.modelDODGE.SPIRIT 62.49966 24.23407 2.579
## make.modelDODGE.STRATUS 23.15683 9.14873 2.531
## make.modelEAGLE.VISION 28.75787 11.17760 2.573
## make.modelEAGLE.VISION TSI 6.63225 33.02050 0.201
## make.modelFORD.CONTOUR 39.08597 8.94157 4.371
## make.modelFORD.CROWN VICTORIA 27.45347 9.18252 2.990
## make.modelFORD.ESCORT 62.18430 8.95804 6.942
## make.modelFORD.FOCUS 77.95125 33.08468 2.356
## make.modelFORD.TAURUS 39.20624 8.78714 4.462
## make.modelFORD.TEMPO 57.47715 12.46621 4.611
## make.modelGEO.METRO 78.94696 10.54549 7.486
## make.modelGEO.PRIZM 54.04597 9.00689 6.001
## make.modelHONDA.ACCORD -5.30500 9.25315 -0.573
## make.modelHONDA.CIVIC 11.64565 9.14550 1.273
## make.modelHYUNDAI.ACCENT 68.56017 9.87369 6.944
## make.modelHYUNDAI.ACCENT GL 26.44021 33.06887 0.800
## make.modelHYUNDAI.ELANTRA 59.32581 9.71897 6.104
## make.modelHYUNDAI.SONATA 40.76738 10.44320 3.904
## make.modelINFINITI.G20 5.18323 10.38870 0.499
## make.modelINFINITI.I30 23.33452 24.14135 0.967
## make.modelINFINITI.J30 -22.87647 14.78141 -1.548
## make.modelINFINITI.Q45 -21.34938 20.29571 -1.052
## make.modelKIA.SEPHIA 68.53003 9.20237 7.447
## make.modelLEXUS.300 -59.28295 33.03832 -1.794
## make.modelLINCOLN.TOWN CAR 7.71553 9.95438 0.775
## make.modelMAZDA.626 20.25437 8.94229 2.265
## make.modelMAZDA.MILLENIA -8.78679 14.78348 -0.594
## make.modelMAZDA.PROTEGE 48.45839 8.96695 5.404
## make.modelMERCURY.GRAND MARQUIS 17.31244 9.30606 1.860
## make.modelMERCURY.MYSTIQUE 42.83178 9.62665 4.449
## make.modelMERCURY.SABLE 41.63974 9.24088 4.506
## make.modelMERCURY.TOPAZ 38.30663 20.39309 1.878
## make.modelMERCURY.TRACER 65.99223 9.36346 7.048
## make.modelMITSUBISHI.GALANT 31.15431 9.15631 3.402
## make.modelMITSUBISHI.MIRAGE 47.98898 9.71985 4.937
## make.modelMITSUBISHI.MIRAGE DE 55.70893 33.10278 1.683
## make.modelNISSAN.ALTIMA 20.31053 8.78908 2.311
## make.modelNISSAN.MAXIMA -1.14425 9.43751 -0.121
## make.modelNISSAN.SENTRA 51.90779 8.77009 5.919
## make.modelNISSAN.STANZA 32.17944 16.63301 1.935
## make.modelOLDSMOBILE.88 42.11710 8.98155 4.689
## make.modelOLDSMOBILE.98 32.46005 9.78136 3.319
## make.modelOLDSMOBILE.ACHIEVA 64.76636 9.35627 6.922
## make.modelOLDSMOBILE.ALERO 31.18099 14.89706 2.093
## make.modelOLDSMOBILE.AURORA -2.15774 13.22959 -0.163
## make.modelOLDSMOBILE.CALAIS 37.49355 33.08870 1.133
## make.modelOLDSMOBILE.CIERA 72.41583 9.30227 7.785
## make.modelOLDSMOBILE.CUTLASS 31.60134 10.91729 2.895
## make.modelOLDSMOBILE.CUTLASS SUPREME 55.61610 8.89335 6.254
## make.modelOLDSMOBILE.INTRIGUE 14.58807 11.34280 1.286
## make.modelPLYMOUTH.ACCLAIM 68.45013 18.20182 3.761
## make.modelPLYMOUTH.BREEZE 36.02857 9.35738 3.850
## make.modelPLYMOUTH.NEON 54.46986 9.22505 5.905
## make.modelPONTIAC.BONNEVILLE 25.01650 8.86388 2.822
## make.modelPONTIAC.GRAND AM 38.40057 8.80003 4.364
## make.modelPONTIAC.GRAND PRIX 41.10952 9.04847 4.543
## make.modelPONTIAC.SUNBIRD 50.17367 20.41677 2.457
## make.modelPONTIAC.SUNFIRE 47.08604 9.54049 4.935
## make.modelSATURN.SL-1 23.78875 33.04565 0.720
## make.modelSUZUKI.ESTEEM 73.37584 10.71938 6.845
## make.modelTOYOTA.AVALON 11.27309 33.04125 0.341
## make.modelTOYOTA.CAMRY -3.49243 8.85994 -0.394
## make.modelTOYOTA.COROLLA 34.60042 8.88235 3.895
## make.modelTOYOTA.TERCEL 59.28293 9.35035 6.340
## s_total.cost 2.59805 0.07512 34.588
## s_mileage 0.13748 0.01448 9.496
## Pr(>|t|)
## (Intercept) < 2e-16 ***
## vehicle.age < 2e-16 ***
## color.setBLUE 0.086716 .
## color.setGOLD 0.861354
## color.setGREEN 0.089519 .
## color.setPURPLE 0.534875
## color.setRED 0.112641
## color.setSILVER 0.860640
## color.setWHITE 0.251467
## stateCA 4.57e-12 ***
## stateFL 0.096666 .
## stateGA < 2e-16 ***
## stateNM 2.09e-11 ***
## stateNV 0.166885
## stateTX 5.27e-16 ***
## stateVA 4.43e-07 ***
## make.modelBUICK.CENTURY 4.59e-10 ***
## make.modelBUICK.LESABRE 7.71e-06 ***
## make.modelBUICK.PARK AVENUE 0.002195 **
## make.modelBUICK.REGAL 9.86e-07 ***
## make.modelBUICK.ROADMASTER 0.034376 *
## make.modelBUICK.SKYLARK 3.36e-13 ***
## make.modelCADILLAC.BROUGHAM RWD 0.520608
## make.modelCADILLAC.CATERA 0.719617
## make.modelCADILLAC.DEVILLE 0.856688
## make.modelCADILLAC.FLEETWOOD RWD 0.367282
## make.modelCADILLAC.SEVILLE 0.130477
## make.modelCHEVROLET.CAPRICE 0.512492
## make.modelCHEVROLET.CAPRICE CLASSIC 0.369453
## make.modelCHEVROLET.CAVALIER 1.78e-09 ***
## make.modelCHEVROLET.CELEBRITY 0.844749
## make.modelCHEVROLET.CORSICA 1.10e-14 ***
## make.modelCHEVROLET.LUMINA 2.58e-07 ***
## make.modelCHEVROLET.MALIBU 0.006626 **
## make.modelCHEVROLET.METRO 3.39e-12 ***
## make.modelCHEVROLET.PRIZM 5.59e-08 ***
## make.modelCHRYSLER.CIRRUS 0.192595
## make.modelCHRYSLER.CONCORDE 0.019416 *
## make.modelCHRYSLER.FIFTH AVENUE 0.666466
## make.modelCHRYSLER.LEBARON 0.232373
## make.modelCHRYSLER.LHS 0.255129
## make.modelCHRYSLER.NEW YORKER 0.153820
## make.modelDAEWOO.LANOS < 2e-16 ***
## make.modelDAEWOO.LEGANZA 0.097471 .
## make.modelDAEWOO.NUBIRA 1.11e-08 ***
## make.modelDAEWOO.NUBIRA SX 0.001804 **
## make.modelDODGE.DYNASTY 0.100267
## make.modelDODGE.INTREPID 0.006943 **
## make.modelDODGE.NEON 1.20e-09 ***
## make.modelDODGE.SPIRIT 0.009927 **
## make.modelDODGE.STRATUS 0.011388 *
## make.modelEAGLE.VISION 0.010106 *
## make.modelEAGLE.VISION TSI 0.840819
## make.modelFORD.CONTOUR 1.25e-05 ***
## make.modelFORD.CROWN VICTORIA 0.002801 **
## make.modelFORD.ESCORT 4.18e-12 ***
## make.modelFORD.FOCUS 0.018492 *
## make.modelFORD.TAURUS 8.24e-06 ***
## make.modelFORD.TEMPO 4.08e-06 ***
## make.modelGEO.METRO 7.85e-14 ***
## make.modelGEO.PRIZM 2.05e-09 ***
## make.modelHONDA.ACCORD 0.566446
## make.modelHONDA.CIVIC 0.202922
## make.modelHYUNDAI.ACCENT 4.12e-12 ***
## make.modelHYUNDAI.ACCENT GL 0.423996
## make.modelHYUNDAI.ELANTRA 1.08e-09 ***
## make.modelHYUNDAI.SONATA 9.55e-05 ***
## make.modelINFINITI.G20 0.617843
## make.modelINFINITI.I30 0.333784
## make.modelINFINITI.J30 0.121746
## make.modelINFINITI.Q45 0.292870
## make.modelKIA.SEPHIA 1.06e-13 ***
## make.modelLEXUS.300 0.072793 .
## make.modelLINCOLN.TOWN CAR 0.438310
## make.modelMAZDA.626 0.023539 *
## make.modelMAZDA.MILLENIA 0.552285
## make.modelMAZDA.PROTEGE 6.70e-08 ***
## make.modelMERCURY.GRAND MARQUIS 0.062874 .
## make.modelMERCURY.MYSTIQUE 8.73e-06 ***
## make.modelMERCURY.SABLE 6.70e-06 ***
## make.modelMERCURY.TOPAZ 0.060362 .
## make.modelMERCURY.TRACER 1.97e-12 ***
## make.modelMITSUBISHI.GALANT 0.000671 ***
## make.modelMITSUBISHI.MIRAGE 8.09e-07 ***
## make.modelMITSUBISHI.MIRAGE DE 0.092433 .
## make.modelNISSAN.ALTIMA 0.020865 *
## make.modelNISSAN.MAXIMA 0.903500
## make.modelNISSAN.SENTRA 3.38e-09 ***
## make.modelNISSAN.STANZA 0.053066 .
## make.modelOLDSMOBILE.88 2.79e-06 ***
## make.modelOLDSMOBILE.98 0.000909 ***
## make.modelOLDSMOBILE.ACHIEVA 4.79e-12 ***
## make.modelOLDSMOBILE.ALERO 0.036372 *
## make.modelOLDSMOBILE.AURORA 0.870445
## make.modelOLDSMOBILE.CALAIS 0.257197
## make.modelOLDSMOBILE.CIERA 7.87e-15 ***
## make.modelOLDSMOBILE.CUTLASS 0.003807 **
## make.modelOLDSMOBILE.CUTLASS SUPREME 4.22e-10 ***
## make.modelOLDSMOBILE.INTRIGUE 0.198443
## make.modelPLYMOUTH.ACCLAIM 0.000171 ***
## make.modelPLYMOUTH.BREEZE 0.000119 ***
## make.modelPLYMOUTH.NEON 3.68e-09 ***
## make.modelPONTIAC.BONNEVILLE 0.004780 **
## make.modelPONTIAC.GRAND AM 1.30e-05 ***
## make.modelPONTIAC.GRAND PRIX 5.62e-06 ***
## make.modelPONTIAC.SUNBIRD 0.014013 *
## make.modelPONTIAC.SUNFIRE 8.16e-07 ***
## make.modelSATURN.SL-1 0.471623
## make.modelSUZUKI.ESTEEM 8.21e-12 ***
## make.modelTOYOTA.AVALON 0.732976
## make.modelTOYOTA.CAMRY 0.693457
## make.modelTOYOTA.COROLLA 9.88e-05 ***
## make.modelTOYOTA.TERCEL 2.42e-10 ***
## s_total.cost < 2e-16 ***
## s_mileage < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 31.87 on 7875 degrees of freedom
## Multiple R-squared: 0.2326, Adjusted R-squared: 0.2215
## F-statistic: 20.94 on 114 and 7875 DF, p-value: < 2.2e-16options(repr.plot.width = 12, repr.plot.height = 4)
plot(calc.relimp(model5, rela=TRUE))
vif(model5)## GVIF Df GVIF^(1/(2*Df))
## vehicle.age 3.015959 1 1.736652
## color.set 1.350101 7 1.021673
## state 1.484111 7 1.028603
## make.model 7.965610 97 1.010754
## s_total.cost 2.003860 1 1.415578
## s_mileage 1.778011 1 1.333421Plot the model and check the residuals plots
options(repr.plot.width = 8, repr.plot.height = 6)
par(mfrow = c(2,2))
plot(model5)
Include the interaction variables
The cost may depend on the vehicle age.
model6 <- lm(lot.sale.days~. -total.cost -mileage -makex - domestic.import -vehicle.age.group -vehicle.type + s_total.cost:vehicle.age, data = train_set[,-1])
summary(model6)##
## Call:
## lm(formula = lot.sale.days ~ . - total.cost - mileage - makex -
## domestic.import - vehicle.age.group - vehicle.type + s_total.cost:vehicle.age,
## data = train_set[, -1])
##
## Residuals:
## Min 1Q Median 3Q Max
## -83.506 -22.243 -7.396 18.415 109.759
##
## Coefficients:
## Estimate Std. Error t value
## (Intercept) -211.01288 17.47553 -12.075
## vehicle.age -0.68873 2.16383 -0.318
## color.setBLUE 2.89433 1.64043 1.764
## color.setGOLD -0.19849 1.77386 -0.112
## color.setGREEN 2.57873 1.55328 1.660
## color.setPURPLE 1.14660 1.74029 0.659
## color.setRED 2.93706 1.86252 1.577
## color.setSILVER 0.52437 1.73243 0.303
## color.setWHITE 1.86155 1.56590 1.189
## stateCA -9.07614 1.28765 -7.049
## stateFL -2.39562 1.33494 -1.795
## stateGA -18.14059 1.56988 -11.555
## stateNM -15.95271 2.37897 -6.706
## stateNV -3.30831 2.34342 -1.412
## stateTX -10.38563 1.26670 -8.199
## stateVA -8.28930 1.69816 -4.881
## make.modelBUICK.CENTURY 59.76193 9.16142 6.523
## make.modelBUICK.LESABRE 41.15459 8.83482 4.658
## make.modelBUICK.PARK AVENUE 30.82124 9.75071 3.161
## make.modelBUICK.REGAL 45.12453 8.92242 5.057
## make.modelBUICK.ROADMASTER 22.66505 10.18536 2.225
## make.modelBUICK.SKYLARK 68.71734 9.20835 7.463
## make.modelCADILLAC.BROUGHAM RWD -19.57610 32.98638 -0.593
## make.modelCADILLAC.CATERA -4.22934 20.32121 -0.208
## make.modelCADILLAC.DEVILLE 2.38959 9.38068 0.255
## make.modelCADILLAC.FLEETWOOD RWD -15.14754 16.62684 -0.911
## make.modelCADILLAC.SEVILLE -20.15669 13.19984 -1.527
## make.modelCHEVROLET.CAPRICE 19.96656 24.12067 0.828
## make.modelCHEVROLET.CAPRICE CLASSIC 10.63619 10.82696 0.982
## make.modelCHEVROLET.CAVALIER 55.73927 8.98583 6.203
## make.modelCHEVROLET.CELEBRITY 22.71970 33.27731 0.683
## make.modelCHEVROLET.CORSICA 75.99066 9.48824 8.009
## make.modelCHEVROLET.LUMINA 46.87895 8.76935 5.346
## make.modelCHEVROLET.MALIBU 27.42758 9.05365 3.029
## make.modelCHEVROLET.METRO 75.48377 10.76722 7.011
## make.modelCHEVROLET.PRIZM 59.45209 10.48789 5.669
## make.modelCHRYSLER.CIRRUS 13.79250 9.64531 1.430
## make.modelCHRYSLER.CONCORDE 23.34052 9.34438 2.498
## make.modelCHRYSLER.FIFTH AVENUE -12.77965 32.99846 -0.387
## make.modelCHRYSLER.LEBARON 50.75901 33.16952 1.530
## make.modelCHRYSLER.LHS 12.35860 9.93845 1.244
## make.modelCHRYSLER.NEW YORKER 19.07508 12.56862 1.518
## make.modelDAEWOO.LANOS 124.36584 11.32980 10.977
## make.modelDAEWOO.LEGANZA 45.06598 20.61837 2.186
## make.modelDAEWOO.NUBIRA 77.47280 12.06584 6.421
## make.modelDAEWOO.NUBIRA SX 103.75263 33.10383 3.134
## make.modelDODGE.DYNASTY 70.33896 33.29544 2.113
## make.modelDODGE.INTREPID 25.79064 8.94209 2.884
## make.modelDODGE.NEON 56.46313 9.06234 6.231
## make.modelDODGE.SPIRIT 65.11088 24.21895 2.688
## make.modelDODGE.STRATUS 25.22744 9.15405 2.756
## make.modelEAGLE.VISION 30.53879 11.17532 2.733
## make.modelEAGLE.VISION TSI 7.76540 32.98941 0.235
## make.modelFORD.CONTOUR 40.64051 8.94106 4.545
## make.modelFORD.CROWN VICTORIA 29.40643 9.18619 3.201
## make.modelFORD.ESCORT 63.33328 8.95377 7.073
## make.modelFORD.FOCUS 84.27203 33.08912 2.547
## make.modelFORD.TAURUS 40.97989 8.78945 4.662
## make.modelFORD.TEMPO 66.60743 12.65619 5.263
## make.modelGEO.METRO 79.32603 10.53559 7.529
## make.modelGEO.PRIZM 55.71902 9.00755 6.186
## make.modelHONDA.ACCORD -5.03490 9.24434 -0.545
## make.modelHONDA.CIVIC 13.02945 9.14293 1.425
## make.modelHYUNDAI.ACCENT 68.92712 9.86445 6.987
## make.modelHYUNDAI.ACCENT GL 27.26822 33.03717 0.825
## make.modelHYUNDAI.ELANTRA 60.39375 9.71305 6.218
## make.modelHYUNDAI.SONATA 42.57494 10.44252 4.077
## make.modelINFINITI.G20 6.51068 10.38371 0.627
## make.modelINFINITI.I30 24.79651 24.12045 1.028
## make.modelINFINITI.J30 -24.19876 14.77056 -1.638
## make.modelINFINITI.Q45 -23.90947 20.28570 -1.179
## make.modelKIA.SEPHIA 69.40668 9.19592 7.548
## make.modelLEXUS.300 -63.38232 33.02152 -1.919
## make.modelLINCOLN.TOWN CAR 7.91431 9.94477 0.796
## make.modelMAZDA.626 21.86547 8.94239 2.445
## make.modelMAZDA.MILLENIA -8.84554 14.76903 -0.599
## make.modelMAZDA.PROTEGE 49.98070 8.96605 5.574
## make.modelMERCURY.GRAND MARQUIS 18.68570 9.30313 2.009
## make.modelMERCURY.MYSTIQUE 44.57026 9.62680 4.630
## make.modelMERCURY.SABLE 43.13866 9.23925 4.669
## make.modelMERCURY.TOPAZ 47.39914 20.49634 2.313
## make.modelMERCURY.TRACER 67.09564 9.35827 7.170
## make.modelMITSUBISHI.GALANT 32.59589 9.15428 3.561
## make.modelMITSUBISHI.MIRAGE 49.76749 9.72026 5.120
## make.modelMITSUBISHI.MIRAGE DE 61.91141 33.10582 1.870
## make.modelNISSAN.ALTIMA 21.80933 8.78828 2.482
## make.modelNISSAN.MAXIMA -1.11015 9.42829 -0.118
## make.modelNISSAN.SENTRA 53.69231 8.77258 6.120
## make.modelNISSAN.STANZA 35.75210 16.64012 2.149
## make.modelOLDSMOBILE.88 43.97155 8.98443 4.894
## make.modelOLDSMOBILE.98 33.57297 9.77565 3.434
## make.modelOLDSMOBILE.ACHIEVA 66.21013 9.35391 7.078
## make.modelOLDSMOBILE.ALERO 35.81753 14.92641 2.400
## make.modelOLDSMOBILE.AURORA -1.59669 13.21738 -0.121
## make.modelOLDSMOBILE.CALAIS 53.32007 33.28626 1.602
## make.modelOLDSMOBILE.CIERA 74.76744 9.31127 8.030
## make.modelOLDSMOBILE.CUTLASS 34.18442 10.92523 3.129
## make.modelOLDSMOBILE.CUTLASS SUPREME 57.02439 8.89145 6.413
## make.modelOLDSMOBILE.INTRIGUE 18.12267 11.36523 1.595
## make.modelPLYMOUTH.ACCLAIM 74.64677 18.24821 4.091
## make.modelPLYMOUTH.BREEZE 37.91738 9.35984 4.051
## make.modelPLYMOUTH.NEON 55.89287 9.22272 6.060
## make.modelPONTIAC.BONNEVILLE 26.76742 8.86575 3.019
## make.modelPONTIAC.GRAND AM 40.20023 8.80263 4.567
## make.modelPONTIAC.GRAND PRIX 43.43426 9.05781 4.795
## make.modelPONTIAC.SUNBIRD 61.01479 20.57150 2.966
## make.modelPONTIAC.SUNFIRE 48.50267 9.53757 5.085
## make.modelSATURN.SL-1 24.54476 33.01387 0.743
## make.modelSUZUKI.ESTEEM 74.10668 10.71042 6.919
## make.modelTOYOTA.AVALON 10.55125 33.00943 0.320
## make.modelTOYOTA.CAMRY -3.32556 8.85138 -0.376
## make.modelTOYOTA.COROLLA 36.36750 8.88438 4.093
## make.modelTOYOTA.TERCEL 61.02958 9.35114 6.526
## s_total.cost 1.87609 0.19332 9.705
## s_mileage 0.13470 0.01448 9.302
## vehicle.age:s_total.cost 0.12700 0.03134 4.052
## Pr(>|t|)
## (Intercept) < 2e-16 ***
## vehicle.age 0.750270
## color.setBLUE 0.077707 .
## color.setGOLD 0.910906
## color.setGREEN 0.096917 .
## color.setPURPLE 0.510009
## color.setRED 0.114853
## color.setSILVER 0.762143
## color.setWHITE 0.234550
## stateCA 1.96e-12 ***
## stateFL 0.072762 .
## stateGA < 2e-16 ***
## stateNM 2.14e-11 ***
## stateNV 0.158064
## stateTX 2.81e-16 ***
## stateVA 1.07e-06 ***
## make.modelBUICK.CENTURY 7.31e-11 ***
## make.modelBUICK.LESABRE 3.24e-06 ***
## make.modelBUICK.PARK AVENUE 0.001579 **
## make.modelBUICK.REGAL 4.35e-07 ***
## make.modelBUICK.ROADMASTER 0.026092 *
## make.modelBUICK.SKYLARK 9.40e-14 ***
## make.modelCADILLAC.BROUGHAM RWD 0.552890
## make.modelCADILLAC.CATERA 0.835137
## make.modelCADILLAC.DEVILLE 0.798934
## make.modelCADILLAC.FLEETWOOD RWD 0.362308
## make.modelCADILLAC.SEVILLE 0.126791
## make.modelCHEVROLET.CAPRICE 0.407821
## make.modelCHEVROLET.CAPRICE CLASSIC 0.325943
## make.modelCHEVROLET.CAVALIER 5.82e-10 ***
## make.modelCHEVROLET.CELEBRITY 0.494792
## make.modelCHEVROLET.CORSICA 1.32e-15 ***
## make.modelCHEVROLET.LUMINA 9.26e-08 ***
## make.modelCHEVROLET.MALIBU 0.002458 **
## make.modelCHEVROLET.METRO 2.57e-12 ***
## make.modelCHEVROLET.PRIZM 1.49e-08 ***
## make.modelCHRYSLER.CIRRUS 0.152765
## make.modelCHRYSLER.CONCORDE 0.012516 *
## make.modelCHRYSLER.FIFTH AVENUE 0.698559
## make.modelCHRYSLER.LEBARON 0.125985
## make.modelCHRYSLER.LHS 0.213715
## make.modelCHRYSLER.NEW YORKER 0.129137
## make.modelDAEWOO.LANOS < 2e-16 ***
## make.modelDAEWOO.LEGANZA 0.028865 *
## make.modelDAEWOO.NUBIRA 1.43e-10 ***
## make.modelDAEWOO.NUBIRA SX 0.001730 **
## make.modelDODGE.DYNASTY 0.034669 *
## make.modelDODGE.INTREPID 0.003935 **
## make.modelDODGE.NEON 4.89e-10 ***
## make.modelDODGE.SPIRIT 0.007194 **
## make.modelDODGE.STRATUS 0.005867 **
## make.modelEAGLE.VISION 0.006296 **
## make.modelEAGLE.VISION TSI 0.813912
## make.modelFORD.CONTOUR 5.57e-06 ***
## make.modelFORD.CROWN VICTORIA 0.001374 **
## make.modelFORD.ESCORT 1.64e-12 ***
## make.modelFORD.FOCUS 0.010890 *
## make.modelFORD.TAURUS 3.18e-06 ***
## make.modelFORD.TEMPO 1.46e-07 ***
## make.modelGEO.METRO 5.67e-14 ***
## make.modelGEO.PRIZM 6.49e-10 ***
## make.modelHONDA.ACCORD 0.586012
## make.modelHONDA.CIVIC 0.154172
## make.modelHYUNDAI.ACCENT 3.03e-12 ***
## make.modelHYUNDAI.ACCENT GL 0.409181
## make.modelHYUNDAI.ELANTRA 5.30e-10 ***
## make.modelHYUNDAI.SONATA 4.61e-05 ***
## make.modelINFINITI.G20 0.530672
## make.modelINFINITI.I30 0.303968
## make.modelINFINITI.J30 0.101397
## make.modelINFINITI.Q45 0.238579
## make.modelKIA.SEPHIA 4.93e-14 ***
## make.modelLEXUS.300 0.054967 .
## make.modelLINCOLN.TOWN CAR 0.426157
## make.modelMAZDA.626 0.014501 *
## make.modelMAZDA.MILLENIA 0.549240
## make.modelMAZDA.PROTEGE 2.57e-08 ***
## make.modelMERCURY.GRAND MARQUIS 0.044620 *
## make.modelMERCURY.MYSTIQUE 3.72e-06 ***
## make.modelMERCURY.SABLE 3.08e-06 ***
## make.modelMERCURY.TOPAZ 0.020772 *
## make.modelMERCURY.TRACER 8.20e-13 ***
## make.modelMITSUBISHI.GALANT 0.000372 ***
## make.modelMITSUBISHI.MIRAGE 3.13e-07 ***
## make.modelMITSUBISHI.MIRAGE DE 0.061506 .
## make.modelNISSAN.ALTIMA 0.013099 *
## make.modelNISSAN.MAXIMA 0.906271
## make.modelNISSAN.SENTRA 9.78e-10 ***
## make.modelNISSAN.STANZA 0.031701 *
## make.modelOLDSMOBILE.88 1.01e-06 ***
## make.modelOLDSMOBILE.98 0.000597 ***
## make.modelOLDSMOBILE.ACHIEVA 1.58e-12 ***
## make.modelOLDSMOBILE.ALERO 0.016436 *
## make.modelOLDSMOBILE.AURORA 0.903851
## make.modelOLDSMOBILE.CALAIS 0.109226
## make.modelOLDSMOBILE.CIERA 1.12e-15 ***
## make.modelOLDSMOBILE.CUTLASS 0.001761 **
## make.modelOLDSMOBILE.CUTLASS SUPREME 1.51e-10 ***
## make.modelOLDSMOBILE.INTRIGUE 0.110848
## make.modelPLYMOUTH.ACCLAIM 4.34e-05 ***
## make.modelPLYMOUTH.BREEZE 5.15e-05 ***
## make.modelPLYMOUTH.NEON 1.42e-09 ***
## make.modelPONTIAC.BONNEVILLE 0.002543 **
## make.modelPONTIAC.GRAND AM 5.03e-06 ***
## make.modelPONTIAC.GRAND PRIX 1.65e-06 ***
## make.modelPONTIAC.SUNBIRD 0.003026 **
## make.modelPONTIAC.SUNFIRE 3.75e-07 ***
## make.modelSATURN.SL-1 0.457220
## make.modelSUZUKI.ESTEEM 4.90e-12 ***
## make.modelTOYOTA.AVALON 0.749247
## make.modelTOYOTA.CAMRY 0.707142
## make.modelTOYOTA.COROLLA 4.29e-05 ***
## make.modelTOYOTA.TERCEL 7.15e-11 ***
## s_total.cost < 2e-16 ***
## s_mileage < 2e-16 ***
## vehicle.age:s_total.cost 5.12e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 31.84 on 7874 degrees of freedom
## Multiple R-squared: 0.2342, Adjusted R-squared: 0.223
## F-statistic: 20.94 on 115 and 7874 DF, p-value: < 2.2e-16options(repr.plot.width = 12, repr.plot.height = 4)
plot(calc.relimp(model6, rela=TRUE))
vif(model6)## GVIF Df GVIF^(1/(2*Df))
## vehicle.age 111.933405 1 10.579858
## color.set 1.354227 7 1.021896
## state 1.494057 7 1.029093
## make.model 10.928460 97 1.012403
## s_total.cost 13.298815 1 3.646754
## s_mileage 1.782028 1 1.334926
## vehicle.age:s_total.cost 101.857927 1 10.092469Remove the vehicle.age since the model instable and vehicle.age become not significant after consider interaction vehicle.age:s_total.cost
model7 <- lm(lot.sale.days~. -total.cost -mileage -makex - domestic.import -vehicle.age.group -vehicle.type + s_total.cost:vehicle.age -vehicle.age, data = train_set[,-1])
summary(model7)##
## Call:
## lm(formula = lot.sale.days ~ . - total.cost - mileage - makex -
## domestic.import - vehicle.age.group - vehicle.type + s_total.cost:vehicle.age -
## vehicle.age, data = train_set[, -1])
##
## Residuals:
## Min 1Q Median 3Q Max
## -83.451 -22.251 -7.387 18.407 109.580
##
## Coefficients:
## Estimate Std. Error t value
## (Intercept) -2.151e+02 1.183e+01 -18.190
## color.setBLUE 2.887e+00 1.640e+00 1.760
## color.setGOLD -2.062e-01 1.774e+00 -0.116
## color.setGREEN 2.583e+00 1.553e+00 1.663
## color.setPURPLE 1.141e+00 1.740e+00 0.656
## color.setRED 2.937e+00 1.862e+00 1.577
## color.setSILVER 5.071e-01 1.731e+00 0.293
## color.setWHITE 1.856e+00 1.566e+00 1.185
## stateCA -9.067e+00 1.287e+00 -7.044
## stateFL -2.384e+00 1.334e+00 -1.787
## stateGA -1.814e+01 1.570e+00 -11.553
## stateNM -1.595e+01 2.379e+00 -6.705
## stateNV -3.311e+00 2.343e+00 -1.413
## stateTX -1.037e+01 1.266e+00 -8.193
## stateVA -8.311e+00 1.697e+00 -4.898
## make.modelBUICK.CENTURY 5.961e+01 9.149e+00 6.516
## make.modelBUICK.LESABRE 4.106e+01 8.830e+00 4.651
## make.modelBUICK.PARK AVENUE 3.077e+01 9.749e+00 3.156
## make.modelBUICK.REGAL 4.506e+01 8.919e+00 5.052
## make.modelBUICK.ROADMASTER 2.259e+01 1.018e+01 2.219
## make.modelBUICK.SKYLARK 6.867e+01 9.207e+00 7.459
## make.modelCADILLAC.BROUGHAM RWD -1.970e+01 3.298e+01 -0.597
## make.modelCADILLAC.CATERA -4.429e+00 2.031e+01 -0.218
## make.modelCADILLAC.DEVILLE 2.345e+00 9.379e+00 0.250
## make.modelCADILLAC.FLEETWOOD RWD -1.513e+01 1.663e+01 -0.910
## make.modelCADILLAC.SEVILLE -2.015e+01 1.320e+01 -1.527
## make.modelCHEVROLET.CAPRICE 1.965e+01 2.410e+01 0.816
## make.modelCHEVROLET.CAPRICE CLASSIC 1.059e+01 1.083e+01 0.978
## make.modelCHEVROLET.CAVALIER 5.570e+01 8.984e+00 6.200
## make.modelCHEVROLET.CELEBRITY 2.144e+01 3.303e+01 0.649
## make.modelCHEVROLET.CORSICA 7.586e+01 9.478e+00 8.003
## make.modelCHEVROLET.LUMINA 4.682e+01 8.767e+00 5.340
## make.modelCHEVROLET.MALIBU 2.728e+01 9.042e+00 3.017
## make.modelCHEVROLET.METRO 7.556e+01 1.076e+01 7.020
## make.modelCHEVROLET.PRIZM 5.935e+01 1.048e+01 5.662
## make.modelCHRYSLER.CIRRUS 1.375e+01 9.644e+00 1.426
## make.modelCHRYSLER.CONCORDE 2.327e+01 9.341e+00 2.491
## make.modelCHRYSLER.FIFTH AVENUE -1.289e+01 3.299e+01 -0.391
## make.modelCHRYSLER.LEBARON 4.991e+01 3.306e+01 1.510
## make.modelCHRYSLER.LHS 1.232e+01 9.937e+00 1.239
## make.modelCHRYSLER.NEW YORKER 1.902e+01 1.257e+01 1.514
## make.modelDAEWOO.LANOS 1.240e+02 1.127e+01 11.003
## make.modelDAEWOO.LEGANZA 4.429e+01 2.047e+01 2.163
## make.modelDAEWOO.NUBIRA 7.682e+01 1.189e+01 6.461
## make.modelDAEWOO.NUBIRA SX 1.038e+02 3.310e+01 3.137
## make.modelDODGE.DYNASTY 6.913e+01 3.308e+01 2.090
## make.modelDODGE.INTREPID 2.572e+01 8.939e+00 2.878
## make.modelDODGE.NEON 5.645e+01 9.062e+00 6.230
## make.modelDODGE.SPIRIT 6.500e+01 2.421e+01 2.684
## make.modelDODGE.STRATUS 2.514e+01 9.150e+00 2.748
## make.modelEAGLE.VISION 3.045e+01 1.117e+01 2.726
## make.modelEAGLE.VISION TSI 7.722e+00 3.299e+01 0.234
## make.modelFORD.CONTOUR 4.060e+01 8.940e+00 4.541
## make.modelFORD.CROWN VICTORIA 2.929e+01 9.178e+00 3.191
## make.modelFORD.ESCORT 6.333e+01 8.953e+00 7.074
## make.modelFORD.FOCUS 8.388e+01 3.306e+01 2.537
## make.modelFORD.TAURUS 4.092e+01 8.787e+00 4.657
## make.modelFORD.TEMPO 6.596e+01 1.249e+01 5.280
## make.modelGEO.METRO 7.939e+01 1.053e+01 7.538
## make.modelGEO.PRIZM 5.565e+01 9.004e+00 6.180
## make.modelHONDA.ACCORD -5.040e+00 9.244e+00 -0.545
## make.modelHONDA.CIVIC 1.295e+01 9.139e+00 1.417
## make.modelHYUNDAI.ACCENT 6.901e+01 9.861e+00 6.998
## make.modelHYUNDAI.ACCENT GL 2.730e+01 3.304e+01 0.826
## make.modelHYUNDAI.ELANTRA 6.041e+01 9.712e+00 6.220
## make.modelHYUNDAI.SONATA 4.252e+01 1.044e+01 4.073
## make.modelINFINITI.G20 6.429e+00 1.038e+01 0.619
## make.modelINFINITI.I30 2.472e+01 2.412e+01 1.025
## make.modelINFINITI.J30 -2.409e+01 1.477e+01 -1.632
## make.modelINFINITI.Q45 -2.372e+01 2.028e+01 -1.170
## make.modelKIA.SEPHIA 6.945e+01 9.194e+00 7.553
## make.modelLEXUS.300 -6.309e+01 3.301e+01 -1.911
## make.modelLINCOLN.TOWN CAR 7.910e+00 9.944e+00 0.795
## make.modelMAZDA.626 2.179e+01 8.939e+00 2.438
## make.modelMAZDA.MILLENIA -8.828e+00 1.477e+01 -0.598
## make.modelMAZDA.PROTEGE 4.993e+01 8.964e+00 5.570
## make.modelMERCURY.GRAND MARQUIS 1.860e+01 9.299e+00 2.001
## make.modelMERCURY.MYSTIQUE 4.450e+01 9.624e+00 4.624
## make.modelMERCURY.SABLE 4.309e+01 9.238e+00 4.665
## make.modelMERCURY.TOPAZ 4.672e+01 2.038e+01 2.292
## make.modelMERCURY.TRACER 6.710e+01 9.358e+00 7.170
## make.modelMITSUBISHI.GALANT 3.254e+01 9.152e+00 3.555
## make.modelMITSUBISHI.MIRAGE 4.972e+01 9.719e+00 5.116
## make.modelMITSUBISHI.MIRAGE DE 6.154e+01 3.308e+01 1.860
## make.modelNISSAN.ALTIMA 2.174e+01 8.785e+00 2.475
## make.modelNISSAN.MAXIMA -1.100e+00 9.428e+00 -0.117
## make.modelNISSAN.SENTRA 5.361e+01 8.769e+00 6.114
## make.modelNISSAN.STANZA 3.549e+01 1.662e+01 2.135
## make.modelOLDSMOBILE.88 4.387e+01 8.978e+00 4.886
## make.modelOLDSMOBILE.98 3.350e+01 9.772e+00 3.428
## make.modelOLDSMOBILE.ACHIEVA 6.618e+01 9.353e+00 7.076
## make.modelOLDSMOBILE.ALERO 3.556e+01 1.490e+01 2.386
## make.modelOLDSMOBILE.AURORA -1.617e+00 1.322e+01 -0.122
## make.modelOLDSMOBILE.CALAIS 5.213e+01 3.307e+01 1.576
## make.modelOLDSMOBILE.CIERA 7.464e+01 9.302e+00 8.024
## make.modelOLDSMOBILE.CUTLASS 3.405e+01 1.092e+01 3.119
## make.modelOLDSMOBILE.CUTLASS SUPREME 5.696e+01 8.888e+00 6.408
## make.modelOLDSMOBILE.INTRIGUE 1.793e+01 1.135e+01 1.580
## make.modelPLYMOUTH.ACCLAIM 7.422e+01 1.820e+01 4.079
## make.modelPLYMOUTH.BREEZE 3.785e+01 9.357e+00 4.045
## make.modelPLYMOUTH.NEON 5.587e+01 9.222e+00 6.058
## make.modelPONTIAC.BONNEVILLE 2.667e+01 8.860e+00 3.011
## make.modelPONTIAC.GRAND AM 4.013e+01 8.799e+00 4.560
## make.modelPONTIAC.GRAND PRIX 4.331e+01 9.049e+00 4.786
## make.modelPONTIAC.SUNBIRD 6.023e+01 2.042e+01 2.949
## make.modelPONTIAC.SUNFIRE 4.848e+01 9.537e+00 5.083
## make.modelSATURN.SL-1 2.456e+01 3.301e+01 0.744
## make.modelSUZUKI.ESTEEM 7.415e+01 1.071e+01 6.924
## make.modelTOYOTA.AVALON 1.061e+01 3.301e+01 0.321
## make.modelTOYOTA.CAMRY -3.322e+00 8.851e+00 -0.375
## make.modelTOYOTA.COROLLA 3.628e+01 8.879e+00 4.086
## make.modelTOYOTA.TERCEL 6.096e+01 9.348e+00 6.521
## s_total.cost 1.934e+00 6.432e-02 30.069
## s_mileage 1.349e-01 1.446e-02 9.329
## vehicle.age:s_total.cost 1.172e-01 5.144e-03 22.776
## Pr(>|t|)
## (Intercept) < 2e-16 ***
## color.setBLUE 0.078437 .
## color.setGOLD 0.907427
## color.setGREEN 0.096278 .
## color.setPURPLE 0.512077
## color.setRED 0.114810
## color.setSILVER 0.769632
## color.setWHITE 0.235978
## stateCA 2.03e-12 ***
## stateFL 0.074053 .
## stateGA < 2e-16 ***
## stateNM 2.15e-11 ***
## stateNV 0.157689
## stateTX 2.94e-16 ***
## stateVA 9.87e-07 ***
## make.modelBUICK.CENTURY 7.67e-11 ***
## make.modelBUICK.LESABRE 3.36e-06 ***
## make.modelBUICK.PARK AVENUE 0.001603 **
## make.modelBUICK.REGAL 4.48e-07 ***
## make.modelBUICK.ROADMASTER 0.026534 *
## make.modelBUICK.SKYLARK 9.66e-14 ***
## make.modelCADILLAC.BROUGHAM RWD 0.550357
## make.modelCADILLAC.CATERA 0.827378
## make.modelCADILLAC.DEVILLE 0.802584
## make.modelCADILLAC.FLEETWOOD RWD 0.362774
## make.modelCADILLAC.SEVILLE 0.126887
## make.modelCHEVROLET.CAPRICE 0.414802
## make.modelCHEVROLET.CAPRICE CLASSIC 0.328101
## make.modelCHEVROLET.CAVALIER 5.95e-10 ***
## make.modelCHEVROLET.CELEBRITY 0.516260
## make.modelCHEVROLET.CORSICA 1.39e-15 ***
## make.modelCHEVROLET.LUMINA 9.54e-08 ***
## make.modelCHEVROLET.MALIBU 0.002557 **
## make.modelCHEVROLET.METRO 2.40e-12 ***
## make.modelCHEVROLET.PRIZM 1.55e-08 ***
## make.modelCHRYSLER.CIRRUS 0.154022
## make.modelCHRYSLER.CONCORDE 0.012761 *
## make.modelCHRYSLER.FIFTH AVENUE 0.696044
## make.modelCHRYSLER.LEBARON 0.131186
## make.modelCHRYSLER.LHS 0.215259
## make.modelCHRYSLER.NEW YORKER 0.130105
## make.modelDAEWOO.LANOS < 2e-16 ***
## make.modelDAEWOO.LEGANZA 0.030540 *
## make.modelDAEWOO.NUBIRA 1.10e-10 ***
## make.modelDAEWOO.NUBIRA SX 0.001714 **
## make.modelDODGE.DYNASTY 0.036645 *
## make.modelDODGE.INTREPID 0.004016 **
## make.modelDODGE.NEON 4.92e-10 ***
## make.modelDODGE.SPIRIT 0.007286 **
## make.modelDODGE.STRATUS 0.006012 **
## make.modelEAGLE.VISION 0.006430 **
## make.modelEAGLE.VISION TSI 0.814920
## make.modelFORD.CONTOUR 5.67e-06 ***
## make.modelFORD.CROWN VICTORIA 0.001424 **
## make.modelFORD.ESCORT 1.64e-12 ***
## make.modelFORD.FOCUS 0.011206 *
## make.modelFORD.TAURUS 3.26e-06 ***
## make.modelFORD.TEMPO 1.32e-07 ***
## make.modelGEO.METRO 5.32e-14 ***
## make.modelGEO.PRIZM 6.72e-10 ***
## make.modelHONDA.ACCORD 0.585586
## make.modelHONDA.CIVIC 0.156414
## make.modelHYUNDAI.ACCENT 2.81e-12 ***
## make.modelHYUNDAI.ACCENT GL 0.408571
## make.modelHYUNDAI.ELANTRA 5.22e-10 ***
## make.modelHYUNDAI.SONATA 4.69e-05 ***
## make.modelINFINITI.G20 0.535666
## make.modelINFINITI.I30 0.305438
## make.modelINFINITI.J30 0.102771
## make.modelINFINITI.Q45 0.242011
## make.modelKIA.SEPHIA 4.72e-14 ***
## make.modelLEXUS.300 0.055978 .
## make.modelLINCOLN.TOWN CAR 0.426363
## make.modelMAZDA.626 0.014799 *
## make.modelMAZDA.MILLENIA 0.549989
## make.modelMAZDA.PROTEGE 2.63e-08 ***
## make.modelMERCURY.GRAND MARQUIS 0.045470 *
## make.modelMERCURY.MYSTIQUE 3.82e-06 ***
## make.modelMERCURY.SABLE 3.14e-06 ***
## make.modelMERCURY.TOPAZ 0.021929 *
## make.modelMERCURY.TRACER 8.15e-13 ***
## make.modelMITSUBISHI.GALANT 0.000380 ***
## make.modelMITSUBISHI.MIRAGE 3.19e-07 ***
## make.modelMITSUBISHI.MIRAGE DE 0.062902 .
## make.modelNISSAN.ALTIMA 0.013359 *
## make.modelNISSAN.MAXIMA 0.907110
## make.modelNISSAN.SENTRA 1.02e-09 ***
## make.modelNISSAN.STANZA 0.032751 *
## make.modelOLDSMOBILE.88 1.05e-06 ***
## make.modelOLDSMOBILE.98 0.000611 ***
## make.modelOLDSMOBILE.ACHIEVA 1.61e-12 ***
## make.modelOLDSMOBILE.ALERO 0.017065 *
## make.modelOLDSMOBILE.AURORA 0.902618
## make.modelOLDSMOBILE.CALAIS 0.115024
## make.modelOLDSMOBILE.CIERA 1.17e-15 ***
## make.modelOLDSMOBILE.CUTLASS 0.001819 **
## make.modelOLDSMOBILE.CUTLASS SUPREME 1.56e-10 ***
## make.modelOLDSMOBILE.INTRIGUE 0.114231
## make.modelPLYMOUTH.ACCLAIM 4.58e-05 ***
## make.modelPLYMOUTH.BREEZE 5.28e-05 ***
## make.modelPLYMOUTH.NEON 1.44e-09 ***
## make.modelPONTIAC.BONNEVILLE 0.002616 **
## make.modelPONTIAC.GRAND AM 5.19e-06 ***
## make.modelPONTIAC.GRAND PRIX 1.73e-06 ***
## make.modelPONTIAC.SUNBIRD 0.003195 **
## make.modelPONTIAC.SUNFIRE 3.80e-07 ***
## make.modelSATURN.SL-1 0.457001
## make.modelSUZUKI.ESTEEM 4.73e-12 ***
## make.modelTOYOTA.AVALON 0.747845
## make.modelTOYOTA.CAMRY 0.707448
## make.modelTOYOTA.COROLLA 4.44e-05 ***
## make.modelTOYOTA.TERCEL 7.41e-11 ***
## s_total.cost < 2e-16 ***
## s_mileage < 2e-16 ***
## vehicle.age:s_total.cost < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 31.84 on 7875 degrees of freedom
## Multiple R-squared: 0.2342, Adjusted R-squared: 0.2231
## F-statistic: 21.13 on 114 and 7875 DF, p-value: < 2.2e-16options(repr.plot.width = 8, repr.plot.height = 6)
par(mfrow = c(2,2))
plot(model7)
#### Check the residual distribution (normal)
model_stat = augment(model7)
options(repr.plot.width = 4, repr.plot.height = 3)
ggplot(model_stat, aes(.resid)) + geom_histogram()
## Test the learned model on test data #### remove the new level in make.model in test_data
test_set <- test_set %>%
dplyr::filter(!make.model %in% c("ACURA.VIGOR","CHRYSLER.FWD","MAZDA.929","MITSUBISHI.DIAMANTE","SATURN.S-SERIES","SATURN.SL"))
predictTest = predict(model7, newdata = test_set)Calculate the error
SSE = sum((test_set$lot.sale.days - predictTest)^2)
SST = sum((test_set$lot.sale.days - mean(sedan_no_outlier$lot.sale.days))^2)
R_sq = 1 - (SSE/SST)
round(R_sq, 4)## [1] 0.2Use the validate data set on learned model
Remove the new level in make.model in test_data
validate_set <- validate_set %>%
dplyr::filter(!make.model %in% c("BUICK.LASABRE","CADILLAC.FLEETWOOD","MITSUBISHI.DIAMANTE","OLDSMOBILE.REGENCY"))
predictValidate = predict(model7, newdata = validate_set)Calculate the error
SSE = sum((validate_set$lot.sale.days - predictValidate)^2)
SST = sum((validate_set$lot.sale.days - mean(sedan_no_outlier$lot.sale.days))^2)
R_sq = 1 - (SSE/SST)
round(R_sq, 4)## [1] 0.1832Run stepAIC to check which variable can be remove
color.set can be remove because it will cause minimum information loss if remove it
stepAIC(model7, direction = 'both')## Start: AIC=55415.01
## lot.sale.days ~ (total.cost + mileage + vehicle.type + domestic.import +
## vehicle.age + vehicle.age.group + color.set + makex + state +
## make.model + s_total.cost + s_mileage) - total.cost - mileage -
## makex - domestic.import - vehicle.age.group - vehicle.type +
## s_total.cost:vehicle.age - vehicle.age
##
## Df Sum of Sq RSS AIC
## - color.set 7 9344 7991388 55410
## <none> 7982044 55415
## - s_mileage 1 88206 8070250 55501
## - state 7 201639 8183684 55600
## - vehicle.age:s_total.cost 1 525780 8507824 55923
## - make.model 97 1633774 9615819 56709
##
## Step: AIC=55410.35
## lot.sale.days ~ state + make.model + s_total.cost + s_mileage +
## vehicle.age:s_total.cost
##
## Df Sum of Sq RSS AIC
## <none> 7991388 55410
## + color.set 7 9344 7982044 55415
## - s_mileage 1 88900 8080288 55497
## - state 7 202899 8194287 55597
## - s_total.cost:vehicle.age 1 530379 8521768 55922
## - make.model 97 1668919 9660308 56732##
## Call:
## lm(formula = lot.sale.days ~ state + make.model + s_total.cost +
## s_mileage + vehicle.age:s_total.cost, data = train_set[,
## -1])
##
## Coefficients:
## (Intercept)
## -213.0951
## stateCA
## -9.0805
## stateFL
## -2.3994
## stateGA
## -18.1865
## stateNM
## -15.8224
## stateNV
## -3.3721
## stateTX
## -10.4261
## stateVA
## -8.3732
## make.modelBUICK.CENTURY
## 59.4340
## make.modelBUICK.LESABRE
## 40.7891
## make.modelBUICK.PARK AVENUE
## 30.4244
## make.modelBUICK.REGAL
## 44.8658
## make.modelBUICK.ROADMASTER
## 22.4721
## make.modelBUICK.SKYLARK
## 68.8921
## make.modelCADILLAC.BROUGHAM RWD
## -18.7573
## make.modelCADILLAC.CATERA
## -4.2568
## make.modelCADILLAC.DEVILLE
## 2.0376
## make.modelCADILLAC.FLEETWOOD RWD
## -15.7952
## make.modelCADILLAC.SEVILLE
## -20.9379
## make.modelCHEVROLET.CAPRICE
## 19.0335
## make.modelCHEVROLET.CAPRICE CLASSIC
## 10.0049
## make.modelCHEVROLET.CAVALIER
## 55.9155
## make.modelCHEVROLET.CELEBRITY
## 19.9226
## make.modelCHEVROLET.CORSICA
## 75.9782
## make.modelCHEVROLET.LUMINA
## 46.7411
## make.modelCHEVROLET.MALIBU
## 26.9620
## make.modelCHEVROLET.METRO
## 75.7114
## make.modelCHEVROLET.PRIZM
## 59.3438
## make.modelCHRYSLER.CIRRUS
## 13.5121
## make.modelCHRYSLER.CONCORDE
## 22.9570
## make.modelCHRYSLER.FIFTH AVENUE
## -14.3401
## make.modelCHRYSLER.LEBARON
## 47.7362
## make.modelCHRYSLER.LHS
## 11.8437
## make.modelCHRYSLER.NEW YORKER
## 18.3807
## make.modelDAEWOO.LANOS
## 124.0871
## make.modelDAEWOO.LEGANZA
## 44.3821
## make.modelDAEWOO.NUBIRA
## 77.0988
## make.modelDAEWOO.NUBIRA SX
## 104.7169
## make.modelDODGE.DYNASTY
## 70.0166
## make.modelDODGE.INTREPID
## 25.6476
## make.modelDODGE.NEON
## 56.5595
## make.modelDODGE.SPIRIT
## 63.3147
## make.modelDODGE.STRATUS
## 24.9009
## make.modelEAGLE.VISION
## 30.4242
## make.modelEAGLE.VISION TSI
## 8.7856
## make.modelFORD.CONTOUR
## 40.5229
## make.modelFORD.CROWN VICTORIA
## 28.9958
## make.modelFORD.ESCORT
## 63.2993
## make.modelFORD.FOCUS
## 82.0608
## make.modelFORD.TAURUS
## 40.8241
## make.modelFORD.TEMPO
## 65.6199
## make.modelGEO.METRO
## 79.7397
## make.modelGEO.PRIZM
## 55.4914
## make.modelHONDA.ACCORD
## -5.5526
## make.modelHONDA.CIVIC
## 12.8838
## make.modelHYUNDAI.ACCENT
## 69.0701
## make.modelHYUNDAI.ACCENT GL
## 28.3959
## make.modelHYUNDAI.ELANTRA
## 60.5107
## make.modelHYUNDAI.SONATA
## 42.5533
## make.modelINFINITI.G20
## 5.9958
## make.modelINFINITI.I30
## 22.8025
## make.modelINFINITI.J30
## -25.4654
## make.modelINFINITI.Q45
## -24.7312
## make.modelKIA.SEPHIA
## 69.3330
## make.modelLEXUS.300
## -65.2422
## make.modelLINCOLN.TOWN CAR
## 7.6393
## make.modelMAZDA.626
## 21.4511
## make.modelMAZDA.MILLENIA
## -9.5960
## make.modelMAZDA.PROTEGE
## 49.6686
## make.modelMERCURY.GRAND MARQUIS
## 18.2177
## make.modelMERCURY.MYSTIQUE
## 44.2820
## make.modelMERCURY.SABLE
## 42.8659
## make.modelMERCURY.TOPAZ
## 46.9016
## make.modelMERCURY.TRACER
## 67.1141
## make.modelMITSUBISHI.GALANT
## 31.9858
## make.modelMITSUBISHI.MIRAGE
## 49.6510
## make.modelMITSUBISHI.MIRAGE DE
## 61.7551
## make.modelNISSAN.ALTIMA
## 21.0784
## make.modelNISSAN.MAXIMA
## -1.5307
## make.modelNISSAN.SENTRA
## 53.2182
## make.modelNISSAN.STANZA
## 35.1217
## make.modelOLDSMOBILE.88
## 43.5856
## make.modelOLDSMOBILE.98
## 33.3897
## make.modelOLDSMOBILE.ACHIEVA
## 66.4938
## make.modelOLDSMOBILE.ALERO
## 36.0529
## make.modelOLDSMOBILE.AURORA
## -1.3005
## make.modelOLDSMOBILE.CALAIS
## 51.9808
## make.modelOLDSMOBILE.CIERA
## 74.6652
## make.modelOLDSMOBILE.CUTLASS
## 33.2527
## make.modelOLDSMOBILE.CUTLASS SUPREME
## 56.9609
## make.modelOLDSMOBILE.INTRIGUE
## 18.0471
## make.modelPLYMOUTH.ACCLAIM
## 73.3060
## make.modelPLYMOUTH.BREEZE
## 37.5901
## make.modelPLYMOUTH.NEON
## 55.9823
## make.modelPONTIAC.BONNEVILLE
## 26.6047
## make.modelPONTIAC.GRAND AM
## 40.4467
## make.modelPONTIAC.GRAND PRIX
## 43.6538
## make.modelPONTIAC.SUNBIRD
## 59.8963
## make.modelPONTIAC.SUNFIRE
## 48.6727
## make.modelSATURN.SL-1
## 23.2676
## make.modelSUZUKI.ESTEEM
## 74.4492
## make.modelTOYOTA.AVALON
## 8.5735
## make.modelTOYOTA.CAMRY
## -3.6804
## make.modelTOYOTA.COROLLA
## 36.1095
## make.modelTOYOTA.TERCEL
## 61.1865
## s_total.cost
## 1.9276
## s_mileage
## 0.1354
## s_total.cost:vehicle.age
## 0.1176Profit/loss analyses & Recommendations
Use the validate_set to predict the profit
If lot.sale.days < 90 ==> Price = 1.2total.cost #### If lot.sale.days >= 90 & < 120 ==> Price = 1.1total.cost
If lot.sale.days >120 ==> Price = 0.8*total.cost
validate_set$predict = round(predictValidate,0)
head(validate_set)## data.set total.cost lot.sale.days mileage vehicle.type domestic.import
## 1 VALIDATE 4944 30 79188 FAMILY.MEDIUM Domestic
## 2 VALIDATE 5541 19 72594 FAMILY.LARGE Domestic
## 3 VALIDATE 4870 39 82492 FAMILY.MEDIUM Import
## 4 VALIDATE 5397 30 71153 FAMILY.MEDIUM Domestic
## 5 VALIDATE 4983 23 93462 FAMILY.LARGE Domestic
## 6 VALIDATE 4430 4 83382 FAMILY.SMALL Domestic
## vehicle.age vehicle.age.group color.set makex state make.model
## 1 5 FIVE GOLD FORD TX FORD.TAURUS
## 2 7 SEVEN+ BLUE BUICK TX BUICK.LESABRE
## 3 6 SIX GREEN MAZDA TX MAZDA.626
## 4 5 FIVE SILVER FORD TX FORD.TAURUS
## 5 6 SIX RED OLDSMOBILE NV OLDSMOBILE.88
## 6 4 FOUR GREEN FORD CA FORD.ESCORT
## s_total.cost s_mileage predict
## 1 70.31358 281.4036 30
## 2 74.43789 269.4327 60
## 3 69.78539 287.2142 22
## 4 73.46428 266.7452 37
## 5 70.59037 305.7156 56
## 6 66.55825 288.7594 41Calculate the cost.
< 90 days, cost1.2 #### >=90 and <120 days, cost1.1
Else cost*0.8
price <- if(validate_set$predict < 90) {
validate_set$total.cost*1.2 } else {
if(validate_set$predict >= 90 & validate_set$predict < 120){
validate_set$total.cost*1.1} else {
validate_set$total.cost*0.8}
}
validate_set$sellingPrice = price
head(validate_set)## data.set total.cost lot.sale.days mileage vehicle.type domestic.import
## 1 VALIDATE 4944 30 79188 FAMILY.MEDIUM Domestic
## 2 VALIDATE 5541 19 72594 FAMILY.LARGE Domestic
## 3 VALIDATE 4870 39 82492 FAMILY.MEDIUM Import
## 4 VALIDATE 5397 30 71153 FAMILY.MEDIUM Domestic
## 5 VALIDATE 4983 23 93462 FAMILY.LARGE Domestic
## 6 VALIDATE 4430 4 83382 FAMILY.SMALL Domestic
## vehicle.age vehicle.age.group color.set makex state make.model
## 1 5 FIVE GOLD FORD TX FORD.TAURUS
## 2 7 SEVEN+ BLUE BUICK TX BUICK.LESABRE
## 3 6 SIX GREEN MAZDA TX MAZDA.626
## 4 5 FIVE SILVER FORD TX FORD.TAURUS
## 5 6 SIX RED OLDSMOBILE NV OLDSMOBILE.88
## 6 4 FOUR GREEN FORD CA FORD.ESCORT
## s_total.cost s_mileage predict sellingPrice
## 1 70.31358 281.4036 30 5932.8
## 2 74.43789 269.4327 60 6649.2
## 3 69.78539 287.2142 22 5844.0
## 4 73.46428 266.7452 37 6476.4
## 5 70.59037 305.7156 56 5979.6
## 6 66.55825 288.7594 41 5316.0Profit/Loss
validate_set$profit.loss = validate_set$sellingPrice - validate_set$total.cost
head(validate_set)## data.set total.cost lot.sale.days mileage vehicle.type domestic.import
## 1 VALIDATE 4944 30 79188 FAMILY.MEDIUM Domestic
## 2 VALIDATE 5541 19 72594 FAMILY.LARGE Domestic
## 3 VALIDATE 4870 39 82492 FAMILY.MEDIUM Import
## 4 VALIDATE 5397 30 71153 FAMILY.MEDIUM Domestic
## 5 VALIDATE 4983 23 93462 FAMILY.LARGE Domestic
## 6 VALIDATE 4430 4 83382 FAMILY.SMALL Domestic
## vehicle.age vehicle.age.group color.set makex state make.model
## 1 5 FIVE GOLD FORD TX FORD.TAURUS
## 2 7 SEVEN+ BLUE BUICK TX BUICK.LESABRE
## 3 6 SIX GREEN MAZDA TX MAZDA.626
## 4 5 FIVE SILVER FORD TX FORD.TAURUS
## 5 6 SIX RED OLDSMOBILE NV OLDSMOBILE.88
## 6 4 FOUR GREEN FORD CA FORD.ESCORT
## s_total.cost s_mileage predict sellingPrice profit.loss
## 1 70.31358 281.4036 30 5932.8 988.8
## 2 74.43789 269.4327 60 6649.2 1108.2
## 3 69.78539 287.2142 22 5844.0 974.0
## 4 73.46428 266.7452 37 6476.4 1079.4
## 5 70.59037 305.7156 56 5979.6 996.6
## 6 66.55825 288.7594 41 5316.0 886.0Total profit/loss
sum(validate_set$profit.loss)## [1] 3903597