A Personalized Context-Aware Recommender System Based on User-Item Preferences

In the digital world, it has become a challenging task to find items that suit users’ persona and fulfill their need. The reason behind this problem is the unprecedented growth of content and product available online. Recommender System (RS) has emerged as a tool which provides personalized results to users as well as suggestion based on its behavior and past history. Collaborative Filtering (CF), the widely used technique, in the field of RS, provides useful recommendations to users based on similar users. Traditional recommendation approaches such as collaborative filtering and content-based filtering, work on two dimensions, i.e., user-item pair. In addition to this “Context used as third dimension”, is also getting popular among researchers. In the present paper, a new method is proposed, i.e., Context-Aware Recommender System by utilizing both item as well as user preferences based on splitting criteria for movie recommendation applications. In this method, first single item is split into two virtual items based on contextual value and a modified dataset is created. Then, the single user is split into two virtual users based on contextual values. Splitting of any user or item is done only if there is a significant difference between two virtual items (users). Further user-based collaborative filtering is used to generate effective recommendations. The results show the effectiveness of proposed scheme in terms of various performance measure criteria using LDOSCOMODA dataset.