bank

analysis.R

library(caret)
library(dplyr)

data = read.csv2("bank-additional-full.csv")

f = y ~ duration + pdays + euribor3m + month + day_of_week + job + default + age + previous + contact + poutcome + nr.employed + cons.price.idx

# Scaling Parameters

SAMPLE = 10
TREES = 100
TUNE = 1

# Stratified Sampling by y

sample = data %>% group_by(y) %>% sample_n(SAMPLE) %>% ungroup

rf = train(f, data = sample, method = "rf", ntrees = TREES, trControl = trainControl(method = "LOOCV", verboseIter = TRUE), verbose = TRUE)

knn = train(f, data = sample, method = "knn", tuneLength = TUNE, trControl = trainControl(method = "LOOCV", verboseIter = TRUE))

# Timing Example

system.time(train(f, data = sample, method = "rf", trControl = trainControl(method = "LOOCV")))