iz

YELP is a crowed-sourced reviewing platform which connects people to local businesses and let people to review the businesses and see the reviews of any business. In this paper, we are focused on the reviews of the restaurants. We aim to predict the business- stars rating for a restaurant from the business attributes, such as the ambience and intimate attributes. For this paper, we investigated 12th Round Yelp Business Data-set which is publicly available on the official YELP website. Through this data we will predict the stars (rating) of a review. The two feature selections, a) Correlation Coefficient and b) Wrapper Method, with the combination of two machine learning   algorithms, a) Random Forest Tree and Decision Tree Model are used.  After the analysis of the performance of each model, the Random Forest Tree algorithm is found to be the best model for predicting the stars/ratings of the reviews. Moreover, we tried to find the subset or combination of business attributes which impacts positively on business stars/ratings.

Keywords: Correlation Coefficient, Decision Tree, Feature Selection, Machine learning, Predicting, Random Forest, Wrapper Method, Yelp Data-set.