The dataset included a set of words and sentences and their label.

Like: the label for "good" and the label of the sentence "the movie was not good".

From my side, I applied only lemmatization and lowered the characrers for preprocessing.