J4 ›› 2011, Vol. 24 ›› Issue (5): 76-80.

• 目录 • 上一篇    下一篇

基于关联图的改进关联规则在推荐系统中的应用

 王林林1, 石冰1, 胡元1,2, 邢海华1   

  1. 1.山东大学计算机科学与技术学院,山东 济南 250101; 2.中国人民解放军77675部队,西藏 林芝 860000
  • 收稿日期:2011-06-15 出版日期:2011-10-20 发布日期:2011-10-20
  • 作者简介:王林林(1986-),女,硕士研究生,主要研究方向为数据仓库与数据挖掘。 Email:wanglinlin1986126@163.com

An application of improved association rules of an association graph in a recommendation system

 WANG Lin-Lin1, SHI Bing1, HU Yuan1,2, XING Hai-Hua1   

  1. 1.School of Computer Science and Engineering, Shandong University, Jinan 250101, China;
    2.77675 Troop, People’s Liberation Army, Linzhi 860000, China
  • Received:2011-06-15 Online:2011-10-20 Published:2011-10-20

摘要:

      提出了推荐模型中的关联规则挖掘方法的改进,给出了自定义的页面权值的定义,并改进了基于关联图的关联规则挖掘算法,将页面权值应用于关联规则的挖掘中。此算法是利用Web日志中经过预处理后得到的数据进行规则挖掘,将处理后的数据应用正态分布函数来得到页面权值。用页面权值重新计算支持度,最后将得到的支持度应用于改进的规则挖掘算法中,形成一种基于权值的关联图的关联规则算法。

关键词: 页面权值, 正态分布, Web日志数据挖掘, 关联规则

Abstract:

        This paper presents an improved association rule mining algorithm for the recommended system, and our definition for the page weights. We improve the association graph based association rules mining algorithm, and apply the page weights to the mining of association rules. This algorithm employs the data acquired after pretreatment to web log to mine the association rules. Page weights are obtained through the processing of such data with a normal distribution function. The algorithm then uses the page weights to recalculate the page support, which is applied to the improved rule mining algorithm. We can therefore acquire page weights based association rule algorithm of an association graph.

Key words: page weight, normal distribution, web log data mining, association rule

中图分类号: 

  • TP311