Recommending sources in news recommender systems

Recommender systems aim to deliver the most suitable item to the user without the manual effort of the user. It is possible to see the applications of recommender systems in a lot of different domains like music, movies, shopping and news. Recommender system development have many challenges. But the dynamic and diverse environment of news domain makes news recommender systems a little bit more challenging than other domains. During the recommendation process of news articles, personalization and analysis of news content plays an important role. But beyond recommending the articles itself, we think that where the news come from is also very important. Different news sources have their own style, view and way of expression and they may give the user a complete, balanced and wide perspective of news stories. In this paper we explain the need for including news sources in news recommendation and propose a news source recommendation method by finding out the implicit relations and similarities between news sources by using semantics and association rules.