A method for ranking news sources, topics and articles

With numerous news available on the Web every day, the needs for effective news ranking algorithms to satisfy users´ requirements are continuously increasing. Though it has attracted lots of interests in the commercial world, little academic research has been done about news ranking. In this paper, we introduce a virtual graph model to describe the properties of news. Based on the model, we propose a ranking algorithm which can fully exploit the reinforcement between news sources, topics and articles. Our ranking algorithm can be processed on line. In the experiments, it shows good results with huge data from EagleRadio, an academic news recommending system.