Abstract
With the increase in number of e-commerce sites, one finds it difficult to choose and buy a product. It is not possible for a person to read hundreds of product reviews from various sources. This problem can be solved by using text summarization. The issue with most text summarizers is that, they summarize the text but they do not tend to preserve the underlying meaning. The aim of this paper is to overcome this problem by developing a abstractive multi document text summarizer which summarizes reviews that are obtained from multiple sources like amazon, trip advisor, etc. and are stored in the form of multiple documents. The implementation is carried out by first cleaning the data and using a graph based approach that considers tagging parts of speech and building edges with weights and then using TF-IDF to find the most important set of words. The final summary is obtained by using maximum weight graph traversal. Evaluation metric is ROUGE.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
