摘要
本文在CF的基础上重点研究推荐系统中的相似兴趣用户的聚类技术。研究重点集中在聚类方法的可伸缩性、聚类复杂形状和数据的有效性,分析了蚂蚁算法的特性和基于蚂蚁聚类的协同过滤推荐系统,并通过实验测试,表明使了利用蚂蚁的聚类特性能降低推荐系统计算量,从而提高系统的伸缩性。
The page research the clustering technique of the alike interest customer recommended system in CF foundation, studying concentrates flexible in clustering method, the shape to clustering and the usefulness of the data, Analysis the characteristic of the ant Algorithm and a collaborative recommends system based on ant clustering algorithm, and pass the experiment test, expressing to make the exploitation choose the clustering characteristic of the ant can lower to the calculation in recommend system, but in...
出处
《微计算机信息》
北大核心
2008年第9期268-270,共3页
Control & Automation
基金
湖南省科技厅研究项目(04ZH6005)
关键词
蚁群算法
协同过滤
推荐系统
AGENT
Ant Clustering Algorithm
Collaborative Filtering
Recommender systems
Agent