gistfile1.txt

library(gbm)
​
library(caret)
library(ranger)
​
setwd("c:/users/Matt/Dropbox/DataScienceBootcamp/Projects/MachineLearning/Santander/")
​
​
train_standardized$response = ifelse(train_standardized$response==0,"NO","YES")
​
params = expand.grid(eta = seq(0.5,1,.1), colsample_bytree= seq(0,1,.5),
                     nrounds=seq(500,1000,250), max_depth = seq(5,10,1),
                     min_child_weight=1,gamma=0)
​
gbmGrid <-  expand.grid(interaction.depth = c(1, 5, 9),
                        n.trees = seq(100,500,100),
                        shrinkage = 0.1,
                        n.minobsinnode = 20)
​
​
​
train.boost = train(x=train_standardized[,-length(train_standardized)], y=train_standardized$response,
                    method = "ranger",
                    metric = "ROC",
                    maximize = FALSE,
                    tuneLength = 5,
                    trControl = trainControl(method="cv",
                                            number=5,
                                            classProbs = TRUE, 
                                            verboseIter = TRUE, 
                                            summaryFunction = twoClassSummary,savePredictions = TRUE))
​
​
#Make test set predictions (after Test data is preprocessed)
​
preds <- predict(train.boost,newdata=test_standardized,type="prob")  
submission <- data.frame(ID=test.id, TARGET=preds[2])
write.csv(submission, "xxx.csv",row.names=F,quote=F)