Research On Hybrid Article Recommendation Approach In Scientific Social Networks

With the rapid development of information technology, Scientific Social Networks (SSNs) have become the fastest and most convenient way for researchers to communicate their research articles. The increasing number of published articles shared via these platforms has brought up the problem of information overload. However, recommendation systems have made it easier for researchers to navigate through the huge amount of information by recommending articles to the researcher.Nowadays, article recommendation approaches can be classified into two major categories:content based filtering and collaborative filtering approaches. The existing content based filtering approach puts emphasis only on keywords while collaborative filtering approach considers only the likes and dislikes information given by a researcher. The fact that the two approaches fail to provide recommendations in specific situations has not been taken into consideration. In addition, standard content based and collaborative filtering approaches do not utilize additional information available on the scientific social networks platforms, in that social information attached on articles by researchers has been largely ignored. This has raised a need for recommendation systems that can make use of the available social information i.e. tags in these platforms. Having realized that there is little research on hybrid recommendation approach particularly with regard to their application in scientific social networks this research proposed a hybrid recommendation system approach that combines content based filtering and collaborative filtering approaches in order to minimize limitations of the individual approaches. This new hybrid approach incorporates social information into content based filtering approach to improve accuracy of recommendation.The proposed hybrid approach was implemented and evaluated using data from CiteULike, a leading scientific social network. Experimental results show that the proposed approach improves quality of recommendation and solves existing problems of content based and collaborative filtering approaches, thus providing a more effective manner to recommend articles on scientific social networks.