Author : Mr. Mohit Dilip Chaudhari and Prof. Mininath Nighot

Now a day’s use of ecommerce is increased. Peoples buy products and post their opinions, suggestions about topic or product. Also before buying a product they may need to go through those reviews before making any decision. This will help them finalize their choice and take a decision whether or not to purchase the product. If the number of reviews is more than going through all reviews will become time consuming process for users. They will not able to interpret all product reviews and might get confused. We are proposing a method based on sentiment analysis than can be used to interpret reviews and summarize of reviews in the user suitable format. It focuses on specific words or attributes that the customer will be interested about product. Product reviews will be classified based on emotions extracted from reviews. We will be using machine learning based approach and neural network based language model for sentiment analysis of product reviews. To process product reviews, we are proposing the use of Map Reduce technique. We proposing a system which efficiently summarizes the reviews posted by customer to help other for deciding about product. This Hadoop based system significantly reduces time needed for the customer for going through intensive process of reading multiple reviews.

Keywords:  Sentiment analysis; Sentiment polarity categorization; Natural language processing; Product reviews