一种面向用户行为的Web聚类模型

A Clustering Model Based on Web Mining for User Behavior

通过对Web日志文件的挖掘,可以发现Web用户在Web上的访问行为和隐藏在Web日志记录中的规律,并发现不同Web用户在Web访问行为上的相似性,以及他们在Web上的访问偏好和消费习惯等知识。该文阐述了用户浏览行为的URL—UserID关联矩阵表示方法,建立了基于URL—UserID关联矩阵的Web页面聚类模型和Web用户聚类模型,为电子商务和其它基于Web信息服务系统的市场营销、客户保持、潜在客户发现和个性化服务等策略的制定提供了科学决策的依据。

Through web log file mining,we can find visiting behaviors of users,regulations hidden in web log records,similarity of behaviors of different users’accessing,preferences on web visiting and consumption habits.In this paper,we expound the URL-UserID associated matrix express method for browsing behaviors of web users,establish the models of web pages clustering and web consumers clustering based on URL-UserID associated matrix. These can provide scientific basis for e-commerce and some web-based information services,such as marketing,contacting customers,discovering potential customers and providing personalized services,etc.