Handy R script to train data in 'training.csv' using random forest and predict output using inputs from 'test.csv'

csvRandomForest.R

trainingCsv <- "./training.csv"
testCsv <- "./test.csv"

trainingData <- read.csv(trainingCsv)
testData <- read.csv(testCsv)

numColumns <- dim(trainingData)[2]
columnNames <- colnames(trainingData)
stopifnot(numColumns == dim(testData)[2] + 1)
stopifnot(columnNames[-numColumns] == colnames(testData))

require(randomForest)

rf <-
  randomForest(
    x = trainingData[,-numColumns], y = as.numeric(unlist(trainingData[outputColumn])), ntree = 100, mtry = 10
  )
outputColumn <- columnNames[numColumns]
testData[outputColumn] <- predict(rf, newdata = testData)
View(testData)
# varImpPlot(rf)