基于新闻推荐的用户兴趣模型研究

摘要

随着互联网的迅猛发展，涌现出一大批新闻类网站，人们也逐渐开始通过网络获取新闻消息，因此针对不同用户推荐个性化的新闻内容将会极大地帮助网站增加用户粘性。为了提高新闻推荐的准确性，建立一种用户综合兴趣模型，首先根据用户浏览网页的行为习惯建立相应的用户稳定兴趣模型；然后根据新闻的时效性和主流性，提出以新闻新鲜度为基础的试探性推荐方法，建立用户的临时兴趣模型；最后，将这两种模型通过加权进行组合以建立用户综合兴趣模型。实验结果证明，提出的方法能从大量最新发布的新闻中推荐最符合用户阅读偏好的特定新闻文章。%With the rapid development of the Internet,a large number of news websites were emerged and people gradually use the Internet to get news,so,introducing the personalized contents according to the users'different requirements will help news websites increase user stickiness.In order to improve the accuracy of news recommendation,this paper establishes a comprehensive user interest model.First,a stable user interest model is established based on user browsing habits.Then,the freshness - based tentative recommendations are described on the basis of news timeliness and mainstream to get the user 's temporary interest model.Finally,these two models are combined to establish a comprehensive user interest model.The experimental results prove that the proposed method can recommend specific news articles which best meets the user's reading preferences from a large number of the latest published news.