关于新闻推荐算法的研究  

Study on Recommendation Algorithm News

作　　者：李乐田 吴林[1] 高永存[1] 

机构地区：[1]中国传媒大学理学院,北京100024

出　　处：《中国传媒大学学报（自然科学版）》2016年第1期40-44,共5页Journal of Communication University of China：Science and Technology

摘　　要：随着互联网高速发展,越来越多的网友依赖于手机、PC获取外界讯息。其中较为广泛关注的当属新闻阅读,用户对新闻推荐的要求也越来越高,除了各大网站推出的头条外,不同用户渴望读到的新闻千差万别,那么如何更好的满足用户对新闻的个性需求就成为新闻推荐的一个难题,本文使用1000名用户的新闻阅读记录,挖掘用户的阅读习惯,发现用户阅读下一条新闻只与上一条有关,而与之前的行为无关,故本文使用马尔可夫算法对这一问题进行研究。With the rapid development of the Internet,more and more users intend to rely on mobile phones,PC access to get external information.The news reading has got more widespread attention,the users＇ requirements of the news recommendation are increasingly high.In addition to the headlines of the websites,different users have different need and interest of different topics.Therefore,it has been a difficulty to satisfy the individual eager of the users.By studying on the record of 1000 users and mining the user＇ s reading habits,this paper found that users read the news related to the last one,has nothing to do with the previous behavior.So we use Markov Model study on this issue.

关 键 词：马尔可夫模型 新闻推荐 关联规则 数据挖掘 

分 类 号：O29[理学—应用数学]