| |
| Detecting user interested topics is one of the most important issues in a recommender system. A topic recommendation model is proposed in this paper, which utilizes the topic network and the social graph to recommend topics that are interesting to the user. The model could be applied to a social Q&A (Question and Answering) system. First, all the topics in the social Q&A system are organized as a topic network. By analyzing the topics that the users have followed and the topic distribution in the topic network, we explore the most relevant topics for users. Secondly, we observe that the topics which attract the majority of our friends or the topics that our best friends are concerned about, might be prospective interesting topics for us. Therefore, we consider these two kinds of scenarios simultaneously to design the topic recommendation algorithm. Based on the data set derived from the Quora website, the experimental results demonstrate that our algorithm outperforms the standard Collaborative Filtering (CF) in the accuracy. |
|
| Keywords:Computerre Application Technology; Commender Systems; Topic Network; Social graph |
|
