Consider the Gini index, classification error, and entropy in a simple classification setting with two classes. Create a single plot that displays each of these quantities as a function of \(ˆpm1\). The x-axis should display \(ˆpm1\), ranging from 0 to 1, and the y-axis should display the value of the Gini index, classification error, and entropy. Hint: In a setting with two classes, \(ˆpm1=1−ˆpm2\). You could make this plot by hand, but it will be much easier to make inR.
p = seq(0, 1, 0.001)
gini.index = 2 * p * (1 - p)
classification.error = 1 - pmax(p, 1 - p)
cross.entropy = - (p * log(p) + (1 - p) * log(1 - p))
matplot(p, cbind(gini.index, classification.error, cross.entropy), col = c("purple", "blue", "yellow"))
In the lab, a classification tree was applied to the Carseats data set after converting Sales into a qualitative response variable. Now we will seek to predict Sales using regression trees and related approaches, treating the response as a quantitative variable.
(a) Split the data set into a training set and a test set.
library(ISLR)
attach = Carseats
set.seed(100 )
train = sample(1:nrow(Carseats), nrow(Carseats) / 2)
Car.train = Carseats[train, ]
Car.test = Carseats[-train,]
(b) Fit a regression tree to the training set. Plot the tree, and interpret the results. What test MSE do you obtain?
library(tree)
regression.tree = tree(Sales~.,data = Car.train)
summary(regression.tree)
##
## Regression tree:
## tree(formula = Sales ~ ., data = Car.train)
## Variables actually used in tree construction:
## [1] "ShelveLoc" "Price" "Age" "CompPrice" "Advertising"
## [6] "Education"
## Number of terminal nodes: 17
## Residual mean deviance: 1.844 = 337.5 / 183
## Distribution of residuals:
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## -3.81700 -0.89580 -0.01857 0.00000 0.92000 2.91800
plot(regression.tree)
text(regression.tree ,pretty =0)
pred.carseats = predict(regression.tree,newdata = Car.test)
mean((pred.carseats - Car.test$Sales)^2)
## [1] 5.395751
The test MSE is 5.395751
(c) Use cross-validation in order to determine the optimal level of tree complexity. Does pruning the tree improve the test MSE?
set.seed(100)
crossval.car = cv.tree(regression.tree)
plot(crossval.car$size, crossval.car$dev, type = "b")
The optimal level of tree complexity is approximately 13.
prune.car = prune.tree(regression.tree, best = 13)
plot(prune.car)
text(prune.car,pretty=0)
pred.carseats2 = predict(prune.car, newdata= Car.test)
mean((pred.carseats2 - Car.test$Sales)^2)
## [1] 5.39241
Prunning the tree decreased the MSE slightly to 5.39241 with an optimal tree complexity of 13.
(d) Use the bagging approach in order to analyze this data. What test MSE do you obtain? Use the importance() function to determine which variables are most important.
library(randomForest)
set.seed(100)
bag.car = randomForest(Sales~.,data=Car.train,mtry = 10, importance = TRUE)
pred.bag = predict(bag.car,newdata=Car.test)
mean((pred.bag-Car.test$Sales)^2)
## [1] 3.232086
importance(bag.car)
## %IncMSE IncNodePurity
## CompPrice 22.03715485 121.684322
## Income -2.06471346 59.203923
## Advertising 12.04860526 68.661153
## Population -0.04479465 50.769726
## Price 50.10510920 344.292522
## ShelveLoc 63.21499092 600.988200
## Age 16.88765951 150.242555
## Education 1.95643155 39.643977
## Urban -0.78701388 4.491141
## US 4.57627564 5.201986
varImpPlot(bag.car)
The MSE is 3.232086. The most important variables are the price and shelving location.
(e) Use random forests to analyze this data. What test MSE do you obtain? Use the importance() function to determine which variables are most important. Describe the effect of m, the number of variables considered at each split, on the error rate obtained.
library(randomForest)
set.seed(100)
randomforest.car = randomForest(Sales~.,data=Car.train,mtry = 3, importance = TRUE)
pred.rf = predict(randomforest.car,newdata=Car.test)
mean((pred.rf-Car.test$Sales)^2)
## [1] 3.514128
The MSE increased to 3.514128 when using random forests.
This problem involves the OJ data set which is part of the ISLR package.
(a) Create a training set containing a random sample of 800 observations, and a test set containing the remaining observations.
library(ISLR)
attach = OJ
set.seed(100)
train = sample(dim(OJ)[1],800)
OJ.train = OJ[train,]
OJ.test = OJ[-train,]
(b) Fit a tree to the training data, with `Purchase1 as the response and the other variables as predictors. Use the \(summary()\) function to produce summary statistics about the tree, and describe the results obtained. What is the training error rate? How many terminal nodes does the tree have?
OJ.tree = tree(Purchase~., data=OJ.train)
summary(OJ.tree)
##
## Classification tree:
## tree(formula = Purchase ~ ., data = OJ.train)
## Variables actually used in tree construction:
## [1] "LoyalCH" "PriceDiff" "ListPriceDiff" "SalePriceMM"
## Number of terminal nodes: 8
## Residual mean deviance: 0.7235 = 573 / 792
## Misclassification error rate: 0.1588 = 127 / 800
The training error rate is 0.1588. The tree has 8 terminal nodes.
(c) Type in the name of the tree object in order to get a detailed text output. Pick one of the terminal nodes, and interpret the information displayed.
OJ.tree
## node), split, n, deviance, yval, (yprob)
## * denotes terminal node
##
## 1) root 800 1070.00 CH ( 0.61000 0.39000 )
## 2) LoyalCH < 0.5036 355 420.20 MM ( 0.27887 0.72113 )
## 4) LoyalCH < 0.280875 167 130.20 MM ( 0.13174 0.86826 )
## 8) LoyalCH < 0.0356415 59 10.14 MM ( 0.01695 0.98305 ) *
## 9) LoyalCH > 0.0356415 108 106.40 MM ( 0.19444 0.80556 ) *
## 5) LoyalCH > 0.280875 188 254.40 MM ( 0.40957 0.59043 )
## 10) PriceDiff < 0.05 81 74.58 MM ( 0.17284 0.82716 ) *
## 11) PriceDiff > 0.05 107 144.90 CH ( 0.58879 0.41121 ) *
## 3) LoyalCH > 0.5036 445 336.80 CH ( 0.87416 0.12584 )
## 6) LoyalCH < 0.764572 180 204.60 CH ( 0.74444 0.25556 )
## 12) ListPriceDiff < 0.18 49 66.27 MM ( 0.40816 0.59184 ) *
## 13) ListPriceDiff > 0.18 131 101.10 CH ( 0.87023 0.12977 )
## 26) SalePriceMM < 2.04 51 59.94 CH ( 0.72549 0.27451 ) *
## 27) SalePriceMM > 2.04 80 25.59 CH ( 0.96250 0.03750 ) *
## 7) LoyalCH > 0.764572 265 85.16 CH ( 0.96226 0.03774 ) *
Node 8 is terminal Split criterion = LoyalCH < 0.0356415 number of observations = 59 deviance = 10.14 overall prediction = Less than 2% of the obs take the value of CH; remaining 98% take value MM
(d) Create a plot of the tree, and interpret the results.
plot(OJ.tree)
text(OJ.tree,pretty=TRUE)
The most important indicator of “Purchase” appears to be “LoyalCH”.
(e) Predict the response on the test data, and produce a confusion matrix comparing the test labels to the predicted test labels. What is the test error rate?
tree.pred = predict(OJ.tree, newdata = OJ.test, type = "class")
table(tree.pred,OJ.test$Purchase)
##
## tree.pred CH MM
## CH 142 36
## MM 23 69
(142+69)/270
## [1] 0.7814815
78% of the test observation are correctly classified so the test error rate is 22%.
(f) Apply the \(cv.tree()\) function to the training set in order to determine the optimal tree size.
cv.OJ = cv.tree(OJ.tree, FUN = prune.misclass)
cv.OJ
## $size
## [1] 8 6 4 2 1
##
## $dev
## [1] 137 137 144 161 312
##
## $k
## [1] -Inf 0.0 4.5 9.5 157.0
##
## $method
## [1] "misclass"
##
## attr(,"class")
## [1] "prune" "tree.sequence"
(g) Produce a plot with tree size on the x-axis and cross-validated classification error rate on the y-axis.
plot(cv.OJ$size,cv.OJ$dev,type='b', xlab = "Tree size", ylab = "Deviance")
(h) Which tree size corresponds to the lowest cross-validated classification error rate?
The 8 node tree is the smallest with the lowest classification error rate.
(i) Produce a pruned tree corresponding to the optimal tree size obtained using cross-validation. If cross-validation does not lead to selection of a pruned tree, then create a pruned tree with five terminal nodes.
prune.OJ = prune.misclass(OJ.tree, best=8)
plot(prune.OJ)
text(prune.OJ,pretty=0)
(j) Compare the training error rates between the pruned and unpruned trees. Which is higher?
summary(OJ.tree)
##
## Classification tree:
## tree(formula = Purchase ~ ., data = OJ.train)
## Variables actually used in tree construction:
## [1] "LoyalCH" "PriceDiff" "ListPriceDiff" "SalePriceMM"
## Number of terminal nodes: 8
## Residual mean deviance: 0.7235 = 573 / 792
## Misclassification error rate: 0.1588 = 127 / 800
summary(prune.OJ)
##
## Classification tree:
## tree(formula = Purchase ~ ., data = OJ.train)
## Variables actually used in tree construction:
## [1] "LoyalCH" "PriceDiff" "ListPriceDiff" "SalePriceMM"
## Number of terminal nodes: 8
## Residual mean deviance: 0.7235 = 573 / 792
## Misclassification error rate: 0.1588 = 127 / 800
The missclassification error rate is 0.1588 for both.
(k) Compare the test error rates between the pruned and unpruned trees. Which is higher?
tree.pred = predict(prune.OJ, newdata = OJ.test, type = "class")
table(tree.pred,OJ.test$Purchase)
##
## tree.pred CH MM
## CH 142 36
## MM 23 69
(142+69)/270
## [1] 0.7814815
The pruned and unpruned have the same test error rate of approximately 22%.