Personalized Recommendation Approach for Academic Literature Using High-Utility Itemset Mining Technique

As the size of digital academic library is increasing enormously, it has become arduous for the researchers to identify the papers of their interest from this repository. This has escalated researcher’s attention toward the implementation of Recommender Systems (RS) in academic literature domain. The content-based and collaborative filtering-based techniques when applied in the academic literature domain, fail in reflecting the researcher’s personalized preferences in terms of recentness, popularity, etc. This article presents a Personalized Recommendation Approach for Academic Literature which is based on High-Utility Itemset Mining (HUIM) Technique. This approach uses the content of the paper along with user’s personalized preference, for making recommendations. Here, we have utilized a highly efficient HUIM algorithm, EFIM, which has been recently introduced in the literature, to mine the papers having higher utility to the user. Experimental evaluation proves that our work satisfies the researcher’s personalized requirements and also outperforms the existing personalized research paper recommender systems in terms of its time and space complexities.