Predicting the Usefulness of Yelp Reviews
The Yelp Dataset Challenge makes a huge set of user, business, and review data publicly available for machine learning projects. They wish to find interesting trends and patterns in all of the data they have accumulated. Our goal is to predict how useful a review will prove to be to users. We can use review upvotes as a metric. This could have immediate applications – many people rely on Yelp to make consumer choices, so predicting the most helpful reviews to display on a page before they have actually been rated would have a serious impact on user experience.
Research Paper Link: Download Paper