摘要
对移动信息服务中面向用户的个性化问题进行了研究,提出了一种基于移动终端浏览系统的、采用改进的K-近邻分类器的个性化推荐算法,对每个移动用户的新闻访问行为序列进行跟踪、记录、建模和分析,将个性化特征归类存储,实现了实时的个性化新闻推荐.仿真实验结果验证了该个性化推荐模型的实用性和可行性,推荐正确率可以达到70%.
Personalized information service is studied for providing high quality mobile information service to mobile users. Based on the mobile off-line reading system, an algorithm of the individuation recommendation service is proposed. The K near neighbors classification method is adjusted so as to trace, log, model and analyze the series of users＇ actions for news browsing. Then the personalized features are classified and saved, and could be used later in the system to provide the real-time individual information recommendation for each user＇s access to news. Simulation experiment result has certified the utility and feasibility of the individual recommendation algorithm. The accuracy of the recommendation can reach 70 %.
出处
《北京邮电大学学报》
EI
CAS
CSCD
北大核心
2006年第6期21-24,共4页
Journal of Beijing University of Posts and Telecommunications
关键词
智能信息处理
个性化信息服务
推荐
intelligent information dispose
individuation information service
recommend
