Co-clustering neighborhood—based collaborative filtering framework using formal concept analysis

In recent times information on the web is growing exponentially and reaching to the point where humans can no longer deal with it manually or with conventional techniques. Recommender systems (RS) act as an effective tool in this situation and are being utilized in a variety of web-based applications and social media platforms. Collaborative filtering (CF) has been researched as one of the most extensively and successfully used techniques in RS. However, CF techniques encounter a significant number of challenges. One such challenge is data sparsity, which has a detrimental impact on the performance of a recommender system. In this paper, a novel CF technique is proposed for generating viable recommendations in sparse environment. The technique is based on co-clustering and optimized by formal concept analysis (FCA). The proposed research work is also focused on providing prospective recommendations ranking based on both local and global similarity index. This work is carried upon using benchmark book-crossing rating dataset. The proposed technique outperforms the existing sparse linear methods (SLIM) and item-based CF (IBCF) techniques in terms of parameters such as hit ratio (HR) and mean reciprocal hit value (MRHV). The results obtained from the research work finds its application in various recommender systems even in the presence of data sparsity. © 2022, The Author(s), under exclusive licence to Bharati Vidyapeeth's Institute of Computer Applications and Management.