prepareTrainData.R

library(caret)

# Load training samples
train <- read.csv("train.csv")

# Extract and convert
trainOut <- data.frame(PassengerId = train$PassengerId,
                       Survived = train$Survived,
                       Pclass = train$Pclass,
                       Sex = as.numeric(train$Sex),
                       Age = train$Age,
                       SibSp = train$SibSp,
                       Parch = train$Parch,
                       Fare = train$Fare,
                       Embarked = as.numeric(train$Embarked))
trainOut <- trainOut[complete.cases(trainOut), ]

# Split training and cross-validation sets
set.seed(12345)
inTrain = createDataPartition(trainOut$PassengerId, p = 2/3, list = FALSE)

trainOutTrain <- trainOut[inTrain, ]
trainOutX <- trainOut[-inTrain,]

inTrain2 = createDataPartition(trainOutX$PassengerId, p = 1/2, list = FALSE)
trainOutCV = trainOutX[inTrain2, ]
trainOutTest = trainOutX[-inTrain2, ]

write.csv(trainOutTrain, file = "train_pure_train.csv", row.names = FALSE)
write.csv(trainOutCV, file = "train_pure_cv.csv", row.names = FALSE)
write.csv(trainOutTest, file = "train_pure_test.csv", row.names = FALSE)
write.csv(trainOut, file = "train_full.csv", row.names = FALSE)

test <- read.csv("test.csv")

testOut <- data.frame(PassengerId = test$PassengerId,
                       Survived = 0,
                       Pclass = test$Pclass,
                       Sex = as.numeric(test$Sex),
                       Age = test$Age,
                       SibSp = test$SibSp,
                       Parch = test$Parch,
                       Fare = test$Fare,
                       Embarked = as.numeric(test$Embarked))

testOut$Age[is.na(testOut$Age)] <- mean(testOut$Age, na.rm = TRUE)
testOut$Fare[is.na(testOut$Fare)] <- mean(testOut$Fare, na.rm = TRUE)

write.csv(testOut, file = "test_full.csv", row.names = FALSE)