You are given a faculty evaluation dataset that categorizes feedback about the corresponding faculty member into 3 levels: good, average and poor.

The training dataset has 3000+ comments while the test dataset has 1300+ comments.