Abstract: Most of the e-commerce applications uses trust reputation trust models, and sellers’ reputation trust scores are computed with the help of the users feedbacks and comments. The “all good reputation” problem is arises in todays systems as the reputation scores are high for sellers so it becomes difficult for buyers to trust the sellers. In this paper, depend on opinions of the buyers openly given in free feedback comments, We propose Comm Trust for trust evaluation by mining tweets, feedbacks, comments. Our main proposition is: (1) We propose a multidimensional trust for computing reputation scores from user feedback comments or tweets. (2) We propose an algorithm for mining feedback comments and opinion mining and by using topic modelling. Most experiments on twitter, eBay and Amazon data demonstrate that Comm Trust can effective for the “all good reputation” issue and rate sellers effectively. This paper gives a brief introduction to this lime lighted topic and explains few of its application, in our information world and also touch topic likes trust score, reputation trust. This paper will also include the algorithms based on topic modelling.

Keywords: E-Commerce, topic modelling, Feedback Mining, Trust Score, Comments, tweets.