个性化新闻推荐技术研究综述

Survey of Research on Personalized News Recommendation Techniques

新闻每时每刻都在发生,阅读新闻已经成为很多人的习惯。新闻媒体众多,网络媒体凭其迅捷性和便利性成为很多人的首选。网络新闻众多导致新闻过载,这就迫切需要个性化的新闻推荐系统,帮助用户快速地找到感兴趣的新闻。伴随着新闻大数据的产生和移动互联网的蓬勃发展,个性化新闻推荐迎来了新的机遇和挑战。首先介绍了个性化新闻推荐的挑战性;然后提出了个性化新闻推荐系统的基本框架,该框架包含新闻建模、用户建模、推荐引擎和用户接口四个模块,并以该框架为基础,分别综述了每个模块的研究进展,列举了现有的个性化新闻推荐系统中四个模块所采用的技术;最后总结了常用数据集、实验方法、评测指标和未来的研究方向。

News happens all the time and many people have developed the habits of reading news. Among numerous news media, network media gets preference for its convenience and celerity. However, too much net-news results in information overload. So it is crucial to develop personalized news recommendation to help users pick up interesting news rapidly. With the growing of big news data and development of mobile internet, there are new chances and challenges in the domain of personalized news recommendation. Firstly, the challenges of personalized news recommendation are introduced. Secondly, an architecture of personalized news recommendation is proposed, which includes news profile, user profile, recommendation engine and user interface. Then based on this architecture,research development of each component is set forth. Thirdly, the methods of existing system of personalized news recommendation according with the architecture are displayed. Lastly, the datasets, evaluation methods, metrics and the possible research directions in the future are concluded.