train<- read.csv("D:/Users/gry7/Downloads/train_LZdllcl.csv")
test<- read.csv("D:/Users/gry7/Downloads/test_2umaH9m.csv")

inputData<- train
# Create Training Data
input_ones <- inputData[which(inputData$is_promoted == 1), ]  # all 1's
input_zeros <- inputData[which(inputData$is_promoted == 0), ]  # all 0's
set.seed(100)  # for repeatability of samples
input_ones_training_rows <- sample(1:nrow(input_ones), 1*nrow(input_ones))  # 1's for training
input_zeros_training_rows <- sample(1:nrow(input_zeros), 5*nrow(input_ones))  # 0's for training. Pick as many 0's as 1's
training_ones <- input_ones[input_ones_training_rows, ]  
training_zeros <- input_zeros[input_zeros_training_rows, ]
trainingData <- rbind(training_ones, training_zeros)  # row bind the 1's and 0's 

test_ones <- input_ones[-input_ones_training_rows, ]
test_zeros <- input_zeros[-input_zeros_training_rows, ]
testData <- rbind(test_ones, test_zeros)  # row bind the 1's and 0's

trainingData <- trainingData[,c('department' , 'education' , 
                                'recruitment_channel' , 'no_of_trainings' , 'age' ,
                                'previous_year_rating' , 'length_of_service' ,
                                'KPIs_met..80.' , 'awards_won.' , 'avg_training_score' ,
                                'is_promoted')]
testData <- testData[,c('department' , 'education' , 
                        'recruitment_channel' , 'no_of_trainings' , 'age' ,
                        'previous_year_rating' , 'length_of_service' ,
                        'KPIs_met..80.' , 'awards_won.' , 'avg_training_score' ,
                        'is_promoted')]


testtest <- test[,c('department' , 'education' , 
                    'recruitment_channel' , 'no_of_trainings' , 'age' ,
                    'previous_year_rating' , 'length_of_service' ,
                    'KPIs_met..80.' , 'awards_won.' , 'avg_training_score' )]


setDT(trainingData) 
setDT(testData)
setDT(testtest)


labels <- trainingData$is_promoted 
ts_label <- testData$is_promoted
new_tr <- model.matrix(~.+0,data = trainingData[,-c("is_promoted"),with=F]) 
new_ts <- model.matrix(~.+0,data = testData[,-c("is_promoted"),with=F])
new_tst <- model.matrix(~.+0,data = testtest)

#convert factor to numeric 
 labels <- as.numeric(labels)
 ts_label <- as.numeric(ts_label)
 
 dtrain <- xgb.DMatrix(data = new_tr,label = labels) 
 dtest <- xgb.DMatrix(data = new_ts,label=ts_label)
 ftest <- xgb.DMatrix(data = new_tst)
 
params <- list(booster = "gbtree", objective = "binary:logistic", 
                eta=0.08, gamma=0.1, max_depth=6, min_child_weight=1, 
                subsample=1, colsample_bytree=1)

xgbcv <- xgb.cv( params = params, data = dtrain, nrounds = 600,
                 nfold = 5, showsd = T, early.stop.round = 20,
                 stratified = T, metrics="logloss",
                 print.every.n = 10, maximize = F)


xgb1 <- xgb.train (params = params, data = dtrain, nrounds = 155, 
                   watchlist = list(val=dtest,train=dtrain), 
                   print.every.n = 10, early.stop.round = 10, maximize = F , 
                   eval_metric = "logloss")
#model prediction
xgbpred <- predict (xgb1,ftest)
xgbpred <- ifelse (xgbpred > 0.5,1,0)

predictions<- data.frame(predictions =xgbpred)
#write.csv(predictions,"D:/Users/gry7/Downloads/predictions.csv",row.names=F)


#confusion matrix
library(caret)
confusionMatrix (as.factor(xgbpred), as.factor(ts_label))


#view variable importance plot
mat <- xgb.importance (feature_names = colnames(new_tr),model = xgb1)
xgb.plot.importance (importance_matrix = mat[1:20])


precision <- posPredValue(as.factor(xgbpred), as.factor(ts_label), positive="1")
recall <- sensitivity(as.factor(xgbpred), as.factor(ts_label), positive="1")

(2 * precision * recall) / (precision + recall)