Science-related Articles Recommendation System from Big Data

Abstract

Under the explosive increase of global data, the term Big data is mainly used to describe enormous datasets. With the availability of increasingly large quantities of digital information, it is becoming more difficult for researchers to extract and find relevant articles pertinent to their interests. In this system, we propose an approach to discover and recommend the desired articles by combining collaborative filtering (CF) with topic modeling. Correlated Topic Model (CTM) is used for modeling topics. Our approach not only considers the interactions between users through collaborative filtering but also learns the properties of items involved through topic modeling to improve recommendation. In order to handle a large dataset, a Big data analytics tool Hadoop is used to perform processing over distributed clusters. The proposed approach learns the accuracy of the recommendation.