FBP Recommendation System through Sentiment Analysis
In present era, online shopping is becoming more vital and common. People are interested in buying the products through online and they also try to know the quality and genuineness of the product through online. Online market provision allow consumers to choose which products to order and allow these online companies to grasp user purchasing behavior. A conceptual model for suggesting and matching products sold online has been already developed. But the model has failed to suggest the feature based best products. It shows the necessity of Recommendation system for online marketing sites to provide feature based product suggestions. This paper deals with construction of FBP Recommendation system for feature based product suggestions based on the user queries. A Natural Language Processing technique with sentiment analysis has been applied to examine the reviews of Amazon mobile product datasets by considering the star ratings, review date, review accommodation score and the review limit. The Naïve Bayes and Support Vector Machine classification algorithms have been applied on these datasets. The performance of these algorithms on Mobile company reviews for camera, battery and value-for-money features have been tested. The average accuracy value of these two algorithms are compared and Support Vector Machine algorithm has proven as best for this application. This FBP Recommendation system can suggest the best company products for the user requested features.