基于工作仓库的OLAM缓存替换算法  

Cache replacement algorithm of OLAM based on work warehouse

作　　者：马治国[1] 郑洪源[1] 丁秋林[1] 

机构地区：[1]南京航空航天大学信息科学与技术学院,南京210016

出　　处：《计算机应用》2009年第1期205-208,共4页journal of Computer Applications

基　　金：国家自然科学基金资助项目(70701018)

摘　　要：联机分析挖掘技术(OLAM)结合了OLAP多维分析的在线性,灵活性和数据挖掘(DM)的深入性的优点,实现了两者的优势互补。传统的OLAM架构在处理海量数据时,数据重复利用率低,执行时间长。针对传统架构中数据处理的耗时性和对数据需求的多样性,提出了一种基于工作仓库的新的OLAM架构,设计了工作仓库中的缓存替换算法。该算法中运用价值度权衡因子对缓存空间进行数据替换判断。在保证算法执行效率的前提下,提高了缓存替换的准确性。同时将该模型应用于某医疗保险决策支持系统(DSS)中,取得了良好的效果。On-Line Analytical Mining （OLAM） combines the on-line, mobility of OLAP and in-depth of Data Mining （ DM）, and it helps integrate the advantages of each other. When the traditional architecture of the OLAM is used to deal with massive data, it often results in a low exploitation rate of repeated data and a long running time. This paper concerned the time consumption during the data-processing of decision analysis and the diversity of data-requirement, and then put forward a new architecture of OLAM based on work warehouse and designed a new algorithm for data cache replacement. This algorithm used the factor of Valueable to judge the data replacement of cache space. On the premise of peak efficiency, it enhanced the accuracy of cache replacement. At the same time, the architecture was applied to the Medical Insurance DSS for Urban Resident and showed good effect.

关 键 词：联机分析挖掘 工作仓库 数据缓存 决策支持系统 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]