摘要
从用户特征描述出发,分析用户兴趣模型的表达机制,提出一种基于用户兴趣的混合模式推荐方法。该方法将内容过滤和协同过滤的预测值进行加权求和,形成最终的综合相似度。实验结果表明,该方法的性能同时优于基于用户协同过滤的推荐方法和基于内容过滤的推荐方法,推荐系统的推荐质量得到显著提高。
Beginning from the description of user's feature and then the analysis of user's interest model, a mixed model of E-Commerce recommender algorithm based on user's interest is presented. Weighted summarizing of the predicted value of the content-based and collaborative filtering was achieved and the final synthesized similarity is found. It has been tested that the performance of this model is superior to the recommender model based on users' collaborative filtering and that on content filtering according to its improvement.
出处
《系统工程》
CSCD
北大核心
2009年第6期68-72,共5页
Systems Engineering
基金
湖南省普通高校教学改革研究项目(2006-191)
湖南省科技厅软科学研究计划项目(04ZH6005)
湖南教育厅科学研究项目(08C234)
关键词
电子商务
推荐系统
协同过滤
混合推荐
E-Commerce
Recommendation Systems
Collabora^tive Filtering
Mixed Filtering Recommendation