利用LMS规则的预取策略  

A Prefetching Strategy Based on LMS Rule

作　　者：班志杰[1,2] 金瑜[2] 

机构地区：[1]内蒙古大学电子信息工程学院 [2]北京理工大学计算机科学技术学院

出　　处：《武汉大学学报（信息科学版）》2009年第8期1004-1007,共4页Geomatics and Information Science of Wuhan University

摘　　要：以往基于All-Kth-Order Markov模型的Web预取策略没有全面考虑访问序列长度、转移概率、网页访问频率、预测准确率和输出概率分布等重要特性。针对这一缺陷,提出了一种基于LMS(least meansquare)规则的预取策略All-Kth-Order LMS。基于日志驱动的实验表明,All-Kth-Order LMS在Web预取性能上具有明显的优势。Existing All-Kth-Order Markov models for Web prefetching only consider one or a few important characteristics such as the length of access sequence, transition probability, page access frequency, prediction precision and output probability distribution. In order to make use of these characteristics, we present a new prefetching strategy based on the LMS （least mean square） rule by extending All-Kth-Order-Markov model. This approach defines a linear function for every state to represent its prediction capabilities and makes use of the maximum entropy to describe a state＇s outgoing probability distribution. In order to improve prediction precision, the function＇s weight values are dynamically updated according to the LMS rule. The trace-driven experiment Shows that our method has a better prefetching performance.

关 键 词：马尔可夫模型 LMS法则 WEB预取 

分 类 号：P208[天文地球—地图制图学与地理信息工程] TP393[天文地球—测绘科学与技术]