PART I - Aspect-level sentiment analysis of Restaurants reviews This project utilizes the data colected from Yelp.com. The reviews of 100 resturants of Montreal are collected and analyzed using NLP techniques. Specifically, it deals with aspect-levelé phrase level senitment analysis. Based on the sentiments of different phrases present in the review, it predicts if the review is postive or negative.
Predicting user ratings based on the sentiment predictions
It gathers all the positive and negative reviews of each restaurant in a dataset. This data set is trained with the sentiments , phrases and 5-star ratings for each restaurant. The final predictions are made by training a model on ratings and reviews analysed using sentiment analysis.