摘要
(minghuay@bit.edu.cn)摘要:为了降低数据规模,并从行为日志中发现更有推荐价值的访问模式,提出了基于用户兴趣特征的数据预处理方法。该方法过滤不具有推荐价值的、用户因偶然发生的短期兴趣而访问网络的行为记录。实验结果表明该方法能够较好地降低数据规模,过滤掉噪音数据,从而减小代理端日志挖掘的复杂度,提高基于Web使用挖掘(WUM)进行个性化推荐的准确度。
To reduce the data scale and find more recommendable access patterns from log file, a new data preprocessing method based on the characteristic of users' interests for Web Usage Mining(WUM) was proposed in this paper. This method filtered out the access records which were caused by users' short-term interests and not recommendable from log file. Experimental results indicate that this method can filter out the noise data so as to reduce the data scale and the complexity of WUM greatly, and enhance the accuracy of WUM - based personalized recommendation.
出处
《计算机应用》
CSCD
北大核心
2006年第10期2393-2394,2397,共3页
journal of Computer Applications
基金
北京理工大学基础研究基金资助项目(0301F18)
关键词
WEB使用挖掘
兴趣品质
兴趣特征
数据预处理
Web Usage Mining(WUM)
interest quality
characteristic of interests
data preprdcessing