Kaggle Amazon employee competition

training = read.csv("train.csv")
testing = read.csv("test.csv")


training$ACTION = as.factor(training$ACTION)


libs = c("caret", "randomForest", "rpart", "kernlab", "glmnet", "gbm",
  	 "doMC")
for (l in libs) {
    if (! require(l, character.only=TRUE) ) {
        install.packages(l)
        library(l, character.only=TRUE) 
    }
}

set.seed(0xDEAD) ## reproducibility!!
trControl = trainControl(method="cv", number = 5)

## 1st hypothesis set : (Simple) decision trees using the 'rpart' package
rf = train(ACTION ~ ., data=training, method="rf", trControl=trControl)
svmRad = train(ACTION ~ ., data=training, method="svmRadial", trControl=trControl)
glmnet = train(ACTION ~ ., data=training, method="glmnet", trControl=trControl)
knn = train(ACTION ~ ., data=training, method="knn", trControl=trControl)
gbm = train(ACTION ~ ., data=training, method="gbm",
			trControl=trControl, verbose=FALSE)

models <- list(svm = svmRad, rf = rf, gbm = gbm, lm = glmnet, knn = knn)
# The best Accuracy results in the training set
sapply(models, function(m) max(m$results$Accuracy))

test_id=testing$id
test_data=testing[, -1]
ACTION = predict(rf, newdata=test_data)
write.table(submit_file, file="rf_1.csv",row.names=FALSE, col.names=TRUE, sep=",")