Extractive review summarization framework for extracted features

Abstract

In the information age, the growth of e-commerce has brought the products’ sale and purchase online and many of the customers prefer to buy it online. To support this preference the users’ reviews of the products plays an important role. So, online merchants wish to take the reviews; experiences of the user, to enhance their business and revenue. Popular and trending products may attract large number of reviews. Further, many of which could be elongated. Extracting useful information with efficiency and accuracy from these so many reviews, of which there are some very long, is a challenging task. This work is an attempt to summarize the customer reviews on products into more useful and shorter version that can help another users’ decision. Reviews available online are crawled for product, each time after extraction, first identification of features of the product will be done and hence polarity will be detected i.e. either a review is positive review or a negative review. After the calculations, summarization of all the features of the product will be generated.