R 머신러닝
heemokang
2021 4 27
R 패키지 설치하고 불러오기
if (! ("caret" %in% rownames(installed.packages())))
{ install.packages("caret") }
base::require("caret")
if (! ("rpart.plot" %in% rownames(installed.packages())))
{ install.packages("rpart.plot") }
base::require("rpart.plot")
if (! ("adabag" %in% rownames(installed.packages())))
{ install.packages("adabag") }
base::require("adabag")
if (! ("randomForest" %in% rownames(installed.packages())))
{ install.packages("randomForest") }
base::require("randomForest")iris data 살펴보기
data("iris")
names(iris)## [1] "Sepal.Length" "Sepal.Width" "Petal.Length" "Petal.Width" "Species"table(iris$Species)##
## setosa versicolor virginica
## 50 50 50ggplot(data=iris,aes(x=Petal.Width,y=Sepal.Width,col=Species)) + geom_point()
데이터 분할하기
70% 데이터 랜덤추출
set.seed(123)
rdata <- createDataPartition(y=iris$Species,p=0.7,list=F)
training <- iris[rdata,]
testing <- iris[-rdata,]iris data 사용한 부스팅 모형
전체 데이터 사용
boosting.adabag <- boosting(Species~.,data=iris, boos = T,mfinal=10)
boosting.adabag$importance## Petal.Length Petal.Width Sepal.Length Sepal.Width
## 65.782868 18.874461 6.518862 8.823809plot(boosting.adabag$trees[[10]])
text(boosting.adabag$trees[[10]])
pred <- predict(boosting.adabag,newdata=iris)
table(pred$class,iris[,5])##
## setosa versicolor virginica
## setosa 50 0 0
## versicolor 0 50 0
## virginica 0 0 50iris data 부스팅 모형
training 데이터 사용
boos는 가중치사용, mfinal는 부스팅 반복횟수
boosting.adabag <- boosting(Species~.,data=training, boos = T,mfinal=10)
boosting.adabag$importance## Petal.Length Petal.Width Sepal.Length Sepal.Width
## 69.385293 21.975015 2.121939 6.517752plot(boosting.adabag$trees[[10]])
text(boosting.adabag$trees[[10]])
pred <- predict(boosting.adabag,newdata=testing)
table(pred$class,testing[,5])##
## setosa versicolor virginica
## setosa 15 0 0
## versicolor 0 14 2
## virginica 0 1 13confusionMatrix(as.factor(pred$class),testing[,5])## Confusion Matrix and Statistics
##
## Reference
## Prediction setosa versicolor virginica
## setosa 15 0 0
## versicolor 0 14 2
## virginica 0 1 13
##
## Overall Statistics
##
## Accuracy : 0.9333
## 95% CI : (0.8173, 0.986)
## No Information Rate : 0.3333
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 0.9
##
## Mcnemar's Test P-Value : NA
##
## Statistics by Class:
##
## Class: setosa Class: versicolor Class: virginica
## Sensitivity 1.0000 0.9333 0.8667
## Specificity 1.0000 0.9333 0.9667
## Pos Pred Value 1.0000 0.8750 0.9286
## Neg Pred Value 1.0000 0.9655 0.9355
## Prevalence 0.3333 0.3333 0.3333
## Detection Rate 0.3333 0.3111 0.2889
## Detection Prevalence 0.3333 0.3556 0.3111
## Balanced Accuracy 1.0000 0.9333 0.9167iris data CART 모형
5-fold cross validation
rdata <- trainControl(method="cv",number = 5)
bagging.rf.kfold <- train(Species~.,data=iris,medhod="treebag",trControl=rdata)
bagging.rf.kfold## Random Forest
##
## 150 samples
## 4 predictor
## 3 classes: 'setosa', 'versicolor', 'virginica'
##
## No pre-processing
## Resampling: Cross-Validated (5 fold)
## Summary of sample sizes: 120, 120, 120, 120, 120
## Resampling results across tuning parameters:
##
## mtry Accuracy Kappa
## 2 0.96 0.94
## 3 0.96 0.94
## 4 0.96 0.94
##
## Accuracy was used to select the optimal model using the largest value.
## The final value used for the model was mtry = 2.pred <- predict(bagging.rf.kfold,newdata=testing)
table(testing$Species,pred)## pred
## setosa versicolor virginica
## setosa 15 0 0
## versicolor 0 15 0
## virginica 0 0 15confusionMatrix(data=pred,reference=testing$Species)## Confusion Matrix and Statistics
##
## Reference
## Prediction setosa versicolor virginica
## setosa 15 0 0
## versicolor 0 15 0
## virginica 0 0 15
##
## Overall Statistics
##
## Accuracy : 1
## 95% CI : (0.9213, 1)
## No Information Rate : 0.3333
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 1
##
## Mcnemar's Test P-Value : NA
##
## Statistics by Class:
##
## Class: setosa Class: versicolor Class: virginica
## Sensitivity 1.0000 1.0000 1.0000
## Specificity 1.0000 1.0000 1.0000
## Pos Pred Value 1.0000 1.0000 1.0000
## Neg Pred Value 1.0000 1.0000 1.0000
## Prevalence 0.3333 0.3333 0.3333
## Detection Rate 0.3333 0.3333 0.3333
## Detection Prevalence 0.3333 0.3333 0.3333
## Balanced Accuracy 1.0000 1.0000 1.0000iris data randomforest 모형
mtry는 tree에서 사용할 변수 갯수
ntree는 tree 총 갯수
rdata <- sample(1:150,120)
training <- iris[rdata,]
testing <- iris[-rdata,]
bagging.rf <- randomForest(Species~.,data=training,mtry=2,ntree=15)
bagging.rf##
## Call:
## randomForest(formula = Species ~ ., data = training, mtry = 2, ntree = 15)
## Type of random forest: classification
## Number of trees: 15
## No. of variables tried at each split: 2
##
## OOB estimate of error rate: 6.67%
## Confusion matrix:
## setosa versicolor virginica class.error
## setosa 41 0 0 0.00000000
## versicolor 0 31 3 0.08823529
## virginica 0 5 40 0.11111111pred <- predict(bagging.rf,newdata=testing)
table(testing$Species,pred)## pred
## setosa versicolor virginica
## setosa 9 0 0
## versicolor 0 15 1
## virginica 0 0 5confusionMatrix(data=pred,reference=testing$Species)## Confusion Matrix and Statistics
##
## Reference
## Prediction setosa versicolor virginica
## setosa 9 0 0
## versicolor 0 15 0
## virginica 0 1 5
##
## Overall Statistics
##
## Accuracy : 0.9667
## 95% CI : (0.8278, 0.9992)
## No Information Rate : 0.5333
## P-Value [Acc > NIR] : 1.759e-07
##
## Kappa : 0.9454
##
## Mcnemar's Test P-Value : NA
##
## Statistics by Class:
##
## Class: setosa Class: versicolor Class: virginica
## Sensitivity 1.0 0.9375 1.0000
## Specificity 1.0 1.0000 0.9600
## Pos Pred Value 1.0 1.0000 0.8333
## Neg Pred Value 1.0 0.9333 1.0000
## Prevalence 0.3 0.5333 0.1667
## Detection Rate 0.3 0.5000 0.1667
## Detection Prevalence 0.3 0.5000 0.2000
## Balanced Accuracy 1.0 0.9688 0.9800