Removes rows from the input data set such that the values in a categorical column are equally distributed.