train.csv is the train data sets.
test.csv is the test data sets.
sample.csv is the sample submission where the second column needs to be filled. The sample.csv is associated with test.csv by the index column.