Feature based opinion summarization of online product reviews
Nilanshi Chauhan, Pardeep Singh · 2017
With the remarkable development of Web 2.0, an amazing growth of the social-media and e-commerce is being witnessed. This expansion of e-commerce has been characterized by the availability of vast number of products online. The websites selling these products allow its customers to express their views freely about a purchased product and these reviews may reach up to hundreds. Hence, it becomes very difficult for the customers to read each and every review and keep track of all the pros and cons of a particular product with respect to all the features that it possesses. The present study targets this problem. Subsequently, a feature based opinion summary of the product reviews is proposed as a solution. The proposed algorithm based on frequent item-set mining gives a set of candidate features for a product. Numerous feature filtering methods like subset filtering, superset filtering and distance based filtering are utilized to get rid of the redundant and very basic features. Finally, sentiment lexicon is used to determine the popularity of the opinion word associated with the extracted feature(s) and a final summary is developed.