Online reviews and their rating play a very important role in information dissemination and are currently the most influencing factor in user decisions. In this project, we have utilized reviews ratings of yelp data challenge[1] to predict how likely a given user would be interested in particular business. The user-user based collaborative filtering algorithm has been used in generating the predictions. The metrics used to evaluate the obtained results are RMSE and accuracy. We have also analyzed the influence of the number of similar neighbors (K) considered on the root mean square error (RMSE) and accuracy of predicted ratings.
Data used: Yelp dataset challenge round 8