基于完全有限前缀的完备日志生成算法  

Generating algorithm for complete log based on complete finite prefix

作　　者：王文星[1,2,3] 闻立杰[1,2,3] 谭士杰[1,2,3] 

机构地区：[1]清华大学软件学院,北京100084 [2]清华大学信息系统安全教育部重点实验室,北京100084 [3]清华大学信息科学与技术国家实验室(筹),北京100084

出　　处：《计算机集成制造系统》2011年第8期1692-1702,共11页Computer Integrated Manufacturing Systems

基　　金：国家自然科学基金资助项目(61003099);国家863计划资助项目(2008AA042301);国家973计划资助项目(2009CB320700);核高基专项资助项目(2010ZX01042-002-002-01)~~

摘　　要：鉴于理想的完备日志很难被获取,针对特定次序关系完备的日志已成为评估挖掘算法表现的重要数据来源。提出了一种新颖的针对特定次序关系完备日志生成算法,基于完全有限前缀计算轨迹,使日志生成过程可控。扩展了完备日志的生成算法,使日志能够针对Δw关系和w关系完备,从而满足α谱系挖掘算法对日志的完备性要求。同时提出了一种日志生成算法的评估框架,并对日志生成算法进行评估。Complete log of specific ordering relations was important data sources for mining algorithm evaluation,but it was difficult to obtain.Aiming at this problem,a new generating algorithm for specific ordering relations complete log was proposed.Generating process could be controlled based on complete finite prefix trace.Relations between Δw and w was completed,and the completeness requirement of α series mining algorithms for log was satisfied by extended complete log generating algorithm.A evaluation framework for log generating algorithm was proposed,and the proposed algorithm was evaluated.

关 键 词：完备日志 展开网 完全有限前缀 α谱系挖掘算法 评估框架 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]