Word Vectors

---

Work with your assigned partner to do sentiment analysis on the IMDB movie data set (reduced)

Do binary classification of sentiment i.e. positive versus negative. The data is already split into training and testing sets but you will need to further divide the data into training/dev/test. Each review is in its own file and the label is the name of the directory it is its in.

Use 50 dimensional GloVe word vectors to represent the reviews rather than a BOW. Use averaging/maxing of the dimensions of the vectors to build a representation for each review. Try different classifiers and report F1 scores. How do word vectors compare to BOW/TF-IDF?

---

1. GFU
2. CSIS
3. J. Walker Orr
4. CSIS 441