← pandyaHomeLab

Text Representation Lab

Preprocessing · stemming vs lemmatization · one-hot, BoW, n-grams, TF-IDF · IMDB reviews

A review, or any text

Try an example:

Prediction failed. Please try again.

1 · Preprocessing pipeline

2 · Tokens: stemming vs lemmatization

3 · Four representations, one classifier

Which representation wins? (IMDB test set)

Training on 25,000 reviews — the table fills in when warm-up finishes.

Worked example: the lecture's four documents