基于带权路径挖掘的缓存选取优化算法

An improved cache optimization Algorithm based on Weighted Path Pattern Mining

阐述了路径模式挖掘的原理,并且针对挖掘对象的特性,把用户浏览路径抽象为特殊的有向图,借鉴有向图的深度遍历算法思想,对AprioriAll算法中生成候选序列的函数做了相应的改进。在此基础上为各路径赋与相应的权值,以表示访问路径的频率,从而在优化缓存内容页面站点选取时不仅仅关注数据挖掘发现的结点序列关联关系,更兼顾到结点的访问频率,改进了缓存页面的选取算法。最后通过实例说明了改进的算法在对服务器缓存选取时的优化作用。

Expressed the principles of path pattern mining details ,according to the characteristics of the mining object,based on the abstraction of the user browse path as a particular directed graph and profits from the depth traversal algorithm thought, improves the function which generating candidate sequences in AprioriAll algorithm.And path evaluated based on visited frequency ,so considered the web visit frequency other than only the relation among the frequent items discovered by data mining.By examples it is showed that the improved algorithm is valid for the server cache optimization.