A content analysis based Recommender System using Categorization of Online Text Resources on Five-Dimensional Human Cognitive System
Rohit Rohit, Tanu Agarwal, Suyash Mani Sharma, Sandeep Gupta, Anil Kumar Singh · 2018 5th IEEE Uttar Pradesh Section International Conference on Electrical, Electronics and Computer Engineering (UPCON) · 2018
A Recommendation System (RS) facilitates customers to purchase items of their choice. A user prefers to get relevant suggestions of his choice. Since, the acceptance of recommendation makes satisfaction to user and it increases the profits of website owner. Therefore, RS has become an important component of e-commerce websites now days. Normally, users give reviews of items as feedback on e-commerce websites. These reviews provide meaningful information about items for other users. The other users use these reviews to decide what he may like from the available options of items. Therefore., Recommender systems (RSs) of e-commerce website largely use the information available in online reviews given by users to generate the recommendations. However., in the influence of social events and latest trends of items., the acceptance of recommended items does not achieve an expected level of satisfaction in user, in spite of inclusion of customer review in computation of recommendation. This paper introduces an innovative approach in RS that uses the information of social events in the computation of recommendations. The approach uses content analysis on online news of social events. The supervised learning method is used in content analysis. The content analysis has two phases namely., news filtration and categorization computation. In news filtration., news is filtered based on their content which may influence the choice of user. The filtered news is then categorized using five-dimensional Human Cognitive System (HCS). The paper has discussed the computation of recommendation using the content analysis of online news using HCS and natural language processing (NLP) on review of products. The final computation of recommendation has enhanced the acceptability by user. It is shown with the help of confusion matrix.