Generating Items Recommendations by Fusing Content and User-Item based Collaborative Filtering

Nowadays e-commerce has spread all over the world. The e-shops are not similar to the physical shops. The e-shops can have hundreds or thousands of items independent of physical boundaries. The information about all these products is available on the Internet. So now customer is overloaded with information. Recommendation System (RS) finds users interest by utilizing implicit or explicit user’s action on the e-commerce website and recommends items that best matches user’s preferences. In this way RS helps in alleviating information overload problem. The two mainstream recommendation systems are Content Based Filtering (CBF) and Collaborative Filtering (CF). The CBF recommends items that have similar characteristics as items used by the user in the past. The CF creates a group of similar users and recommends items to the target user which is preferred in his group. Another variant of collaborative filtering finds and recommends items similar to items rated by the user, it is known as Item based Collaborative Filtering. The CF usually creates a very big recommendation list but users like only small number of items as recommendations. The proposed approach has merged the content and user-item based collaborative filtering and created a single RS that generates relatively small number of recommendations. The main contribution of the paper is to curtail the size of the recommendation list and helps in placing good items in relatively smaller list of recommended items. The experimental results show that the proposed approach surpasses other traditional benchmark recommendation methods.