Abstract:

It is a real challenge to recommend news for users according to their interests from the existing mass information on the Internet. The traditional news propagation way, such as news on the portal sites, does not take the users’ personal needs into consideration, which results in the low accuracy of the recommended news for users. In this article we propose a news sharing model based on the collaboration of users. More exactly, in this model, implicit ratings that the users give to the page are modeled from the users’ behavior data, which are collected from the web browser plugin, and then the news recommendations are shared among similar users based on these ratings through the algorithm of collaborative filtering. The method’s availability is confirmed through a prototypic implementation of the model and the performance of the recommendation algorithm is also analyzed.

Key words: browsing behavior;user interests;news recommendation;collaborative filtering