A Paper Recommendation System Based on User Interest and Citations

Abstract

It is difficult for researchers to access useful scientific articles from an excessive number of articles. This difficulty has led to article recommendation systems. Many studies have been conducted on the article recommendation. Article recommendation studies in the literature generally do not take into account the interest of users. The innovation of this study is to propose research articles by using bibliography and creating a user profile. The proposed approach aims to personalize scientific recommendations based on article-bibliographic relationships and user profile. In our study, candidate recommendation articles were clustered by K-Means clustering method according to their bibliographies. In our study, K-NN algorithm is used to determine which cluster is closer to the user's profile. Once the cluster to which the user belongs is determined, other articles in the cluster are recommended to the user. Experimental results show that the method is successful. © 2019 IEEE.