Yelp-sentiment-analysis

Sentiment analysis of yelp data using deep learning. The yelp reviews data can be downloaded from Here.

The notebook 'preprocess_data.ipynb' pre-processes the yelp dataset and extracts 100,000 positive, neutral and negative reviews.

The notebook 'yelp_sentiment_analysis.ipynb' does two tasks:

1. 70-30 split the 300,000 reviews (proportional wrt classes) into training and test sets and then uses keras to train and test a deep learning model.
2. 70-30 split the 200,000 reviews (positive and negative reviews) into training and test sets (proportional wrt classes) and then uses keras to train and test a deep learning model.