中文摘要：

提出一种基于客户端的网页预取模型。此模型以WPC方法为理论核心，WPC方法以客户端用户访问日志数据为依据，从中挖掘出Web用户的访问模式再通过提出的模式匹配算法进行网页预测．还提出基于Agent的Web预取系统具体应用WPC方法进行网页预取决策。此模型避免或简化了一般的网页预取模型对Web日志进行的复杂的预处理的一些步骤，其中包括数据净化，用户识别，用户会话识别和事务识别等，从而提高了预取效率。实验结果表明该模型能够达到较为理想的预取效果，具有较高的实用价值。

英文摘要：

A client-based pre-fetching model is presented in this paper. The kernel of this model is the WPC method which predicts future requests based on pattern mining and matching,and the patterns are mined from the data in the Web access logs of client.And a Web Pre-Fetching system based on Agent is presented to make pre-fetching decision by utilizing the WPC method.This model avoids or predigests the steps i.e.data cleaning,user identification,session identification,and transaction identification which are necessary for generic Web pre-fetching models,improving the efficiency.The result shows that the model can achieve a relatively perfect pre-fetching effect.