Skip to content

bhparijat/Yelp-reviews-PyTorch-and-Pyspark

Repository files navigation

Sentiment Analysis of Yelp Reviews

The Yelp dataset contains reviews of restaturants listed on the website. Each review is marked with a flag such as cool, funny, sad and given a rank on the scale of 5. I trained two models to predict whether each sentiment can be classified as positive or negative. The two models are:

1. TF-IDF and Logistic Regression with PySpark
2. Word Embeddings neural network model implemented in PyTorch