A Trust Based Collaborative Filtering Recommendation System using Deep Neural Network

Abstract

Recommender system (RS) are a type of suggestion to the information overload problem suffered by user of websites that allow the rating of particular item. RSs are one of the most successful and widespread applications of machine learning technologies in E-commerce. These techniques are used to predict the rating that one individual will give to an item or social entity. It uses the opinions of members of a community to help individuals in that community to identify the information most likely to be interesting to them or relevant to their needs. These systems use the similarity between the user and recommenders or between the items to form the recommendation list for the user. These preferences are being predicted using different approaches, namely content-based approach, collaborative filtering approach, etc. The movie RS are one of the most efficient, useful, and widespread applications for individual to watch movie with minimum decision time. Many attempts made by the researchers to solve these issues like watching movie, purchasing book etc., through RS, whereas most of the study fails to address cold start problem, data sparsity and malicious attacks.