1. Type the service category you are interested in.
2. Type the location you are at.
3. Choose the most important service feature for you.
4. Submit the request.
5. The recommendation list will tell you the true opinions inside the reviews.
A Latent Dirichlet allocation(LDA) model was developed from over 300000 restaurant reviews provided by Yelp open dataset. Reviews were classified as composition of seven restaurant related topics. Each topic was interpreted base on the most important words defining the topic class.
Bag of words model was developed Multinomial Naive Bayes classifier for positive and negative sentiment of review regarding a specific topic.