ExOpSum: An Extractive Opinion Summarization methodology based on aspect-sentence-review ranking

Abstract

With the evolving web technology, anything you need can be purchased online. With these purchases, the feel of the purchase is also recorded in the web portals in the form of user generated comments and feedbacks. The unlimited pour of opinions on the web has now paved way for automated sentiment analysis of opinionated text. Furthermore, in this fast-paced world, humans don’t find time to stand and stare, hence an automated opinion summary is the need of the hour. Summary should include highly ranked aspects, highly ranked sentences and highly ranked reviews. Also, the emotional touch of customers can be deliberately seen only through extractive summarization, which is the core objective of our work. The mathematical computational model designed for our work has proven to work better than the existing work on extractive summarization.