一种基于Web日志挖掘的用户偏爱度度量方法

A method to measure users favoritism based on Web logs mining

分析了用户访问Web站点的浏览日志,度量用户的浏览行为.实验从实际获得的Web日志着手,进行Web日志的挖掘,提取用户浏览Web的行为特性数据.通过时间阈值进行会话的划分,选取合适的数据预处理,归一化后生成数据模式向量,引入人工神经网络中的自组织特征映射(SOM)模型,对用户访问倾向聚类,对用户浏览的偏爱度进行度量,为Web站点的进化提供依据.

Having analyzed browsing logs of Web site accessed by users, the users＇ browsing behaviors can be measured. The experiment obtained the actual acquired Web logs and dealed with Web logs mining, from which the characteristic data of users＇ browsing Web site behavior are extracted. Based on these data, the sessions can be divided by time threshold. After that, the interesting data are initially prepared and then normalized to generate pattern vectors. In order to measure the users favoritism, the Self-Organizing Maps（SOM） model of neural network was introduced to cluster the tendency of users＇ accessing. That provides the effective decision-making reference for Web site evolution.