In today’s competitive business world, the numbers of ecommerce companies are increasing day to day and also huge numbers of products are emerging into the market. When customers want to buy the products they generally just see a numerical rating of the product items and then purchase the items. Further came to know that the products are not good. The current rating system involves direct rating, collaborative rating and content based rating based on the transactions. Based on the rating system, the users are motivating by way of seeing the quality, performance and reviews, but they are missing to see the product features which are available in the products. This may not be taken into consideration, the sentiments expressed on the product unlike few of which consider the reviews and their performance. The sentiment analysis on the reviews does not consider various features available in the existing products. In the research approach the sentiment analysis will be performing per review and per feature. Also the feature based on sentiment analysis algorithm is executed for 3 different scenarios namely single Feature, Multiple Features and No- Feature. For No-feature it can be consider the computation of frequency on each token in the searched query. Finally, graphs are generated, based on each feature which products are the best. For collection of reviews the approach considers an offline review within the application as well as real time reviews for any kind of ecommerce web site by using Web Crawler algorithm. The implementation makes use of latest technology stack namely spring framework for the backend and Ext JS Framework for the front end.
There are no reviews yet.