基于流程日志挖掘的工作流质量分析系统WfQAS的设计与实现
|
摘要
工作流管理系统(WfMS)的主要目标是通过合理地调用和分配有关的信息及人力资源来协调业务流程过程中的各个活动，以促使业务目标的高效实现。为了提高工作流的性能和服务质量，智能化的工作流管理系统应具有分析和优化流程质量的能力。流程日志(Process Logs)跟踪记录了所有流程执行过程的详细信息，因而是分析工作流服务质量的重要依据。
本文意在通过对这些流程日志的分析和挖掘，开发出一个基于工作流日志挖掘技术的流程质量分析系统WfQAS(Workflow Quality Analysis System，以下简称WfQAS)，由此企业决策者可随时获得对所有流程执行质量的评估和分析情况，掌握工作流执行动态。WfQAS主要采用数据仓库和数据挖掘技术，两者相结合提取大量日志中有用的“知识”，以分析流程执行质量和导致性能降低的原因，据此可采取相应措施来优化流程。
WfQAS包括流程日志数据仓库信息决策模块、行为模式关联规则挖掘模块和流程异常分类挖掘分析模块三大部分。
流程日志数据仓库信息决策模块，以巨量的原始流程日志为基础，通过引入事实表和维表，建立相应的数据仓库，并在其上使用OLAP技术进行各种决策统计，可以获得对影响流程质量不同角度的剖析，包括流程模型定义、相关的时间因素、业务数据以及作为流程参与者的各种资源信息等。
行为模式关联规则挖掘模块中，提出了“行为模式”概念；所谓行为，通常是用户“感兴趣”的，也就是他们所关心的、往往与工作流性能相关的流程实例执行状况或结果；WfQAS将流程实例执行过程中表现出的属性定义为某种“行为”，以模板形式保存；利用数据挖掘中的关联规则算法-Apriori，提取出相关联的行为模式组合，分析这些行为“相关”的原因，对导致流程效率降低的行为组合采取对策，提出解决方案。
流程异常分类挖掘分析模块中，提出流程“异常”概念；所谓“异常”，是流程质量降低的行为表观，WfQAS对常见的异常情况进行了预先定义。决策者可以选择所要分析的流程异常类型以及感兴趣的流程行为属性(如流程启动者)，WfQAS将对当前要分析的流程进行数据预处理，然后利用决策树分类挖掘算法，获取导致流程异常的原因，决策者可据此对症下药，解决流程质量的瓶颈。
To improve the performance and service quality of workflows, Intelligent WfMS should provide the ability of Process Quality Analysis. WfMS logs every event that occurs during processes execution. Therefore, workflow logs include a significant amount of information that can be used to analyze process executions, understand the causes of high- and low-quality process executions. In this thesis, I develop a Workflow Quality Analysis System (WfQAS), which deploy data warehousing and mining techniques, to "mine" useful "knowledge" from a large volume of workflow logs for analyzing the reasons of the "good" and "bad" of processes.
WfQAS is comprised of data warehouse model on OLAP technology, correlations among behaviors' mining model and analysis of process exceptions model.
Data warehouse model based on OLAP technology set up a data warehouse by using the huge process logs, on which we can use OLAP, and master the quality of some process from different aspects such as times and resources as participants.
Correlations among behaviors' mining model, based on the predefined Behavior Patterns, which are usually related to quality of processes. WfQAS perform algorithm Apriori to mine correlations among behaviors.
Analysis of process exceptions model, analyze exceptions i.e., of deviations from the desired or acceptable behavior, by decision tree classification algorithm.
本文意在通过对这些流程日志的分析和挖掘，开发出一个基于工作流日志挖掘技术的流程质量分析系统WfQAS(Workflow Quality Analysis System，以下简称WfQAS)，由此企业决策者可随时获得对所有流程执行质量的评估和分析情况，掌握工作流执行动态。WfQAS主要采用数据仓库和数据挖掘技术，两者相结合提取大量日志中有用的“知识”，以分析流程执行质量和导致性能降低的原因，据此可采取相应措施来优化流程。
WfQAS包括流程日志数据仓库信息决策模块、行为模式关联规则挖掘模块和流程异常分类挖掘分析模块三大部分。
流程日志数据仓库信息决策模块，以巨量的原始流程日志为基础，通过引入事实表和维表，建立相应的数据仓库，并在其上使用OLAP技术进行各种决策统计，可以获得对影响流程质量不同角度的剖析，包括流程模型定义、相关的时间因素、业务数据以及作为流程参与者的各种资源信息等。
行为模式关联规则挖掘模块中，提出了“行为模式”概念；所谓行为，通常是用户“感兴趣”的，也就是他们所关心的、往往与工作流性能相关的流程实例执行状况或结果；WfQAS将流程实例执行过程中表现出的属性定义为某种“行为”，以模板形式保存；利用数据挖掘中的关联规则算法-Apriori，提取出相关联的行为模式组合，分析这些行为“相关”的原因，对导致流程效率降低的行为组合采取对策，提出解决方案。
流程异常分类挖掘分析模块中，提出流程“异常”概念；所谓“异常”，是流程质量降低的行为表观，WfQAS对常见的异常情况进行了预先定义。决策者可以选择所要分析的流程异常类型以及感兴趣的流程行为属性(如流程启动者)，WfQAS将对当前要分析的流程进行数据预处理，然后利用决策树分类挖掘算法，获取导致流程异常的原因，决策者可据此对症下药，解决流程质量的瓶颈。
To improve the performance and service quality of workflows, Intelligent WfMS should provide the ability of Process Quality Analysis. WfMS logs every event that occurs during processes execution. Therefore, workflow logs include a significant amount of information that can be used to analyze process executions, understand the causes of high- and low-quality process executions. In this thesis, I develop a Workflow Quality Analysis System (WfQAS), which deploy data warehousing and mining techniques, to "mine" useful "knowledge" from a large volume of workflow logs for analyzing the reasons of the "good" and "bad" of processes.
WfQAS is comprised of data warehouse model on OLAP technology, correlations among behaviors' mining model and analysis of process exceptions model.
Data warehouse model based on OLAP technology set up a data warehouse by using the huge process logs, on which we can use OLAP, and master the quality of some process from different aspects such as times and resources as participants.
Correlations among behaviors' mining model, based on the predefined Behavior Patterns, which are usually related to quality of processes. WfQAS perform algorithm Apriori to mine correlations among behaviors.
Analysis of process exceptions model, analyze exceptions i.e., of deviations from the desired or acceptable behavior, by decision tree classification algorithm.
引文
[1] W．H．Inmon著，《数据仓库》，机械工业出版社，2001．7
[2] Jiawei Han，Micheline Kamber著，《数据挖掘概念与理论》，机械工业出版社，2001．8
[3] Michael Corey等著，《Oracle8i数据仓库》，机械工业出版社，2002
[4] 飞思科技产品研发中心 编著，《Oracle9i数据仓库构建技术》，电子工业出版社，2003
[5] 谈竹奎 仇莉编著，《Oracle9i数据库管理员实用技术指南》，中国铁道出版社，2003
[6] Bjarki Holm，John Carnell等著，《Oracle9i Java程序设计》，清华大学出版社，2002
[7] Bruce Eckel著，《Thinking in Java》second Edition，机械工业出版社，2002
[8] 范玉顺主编，《工作流管理技术基础—实现企业业务过程重组、过程管理与过程自动化的核心技术》，清华大学出版社、施普林格出版社，2001
[9] 罗海滨、范玉顺、吴澄，工作流技术综述，《软件学报》，2000．11
[10] 李建强、范玉顺，基于Petri网的工作流模型性能分析方法，计算机应用，2001．6
[11] 肖晶、黄国兴等，基于流程日志的工作流性能分析和优化，《微型电脑应用》(已被录用)
[12] The Workflow Management Coalition (WfMC). http://www.wfmc.org/.Workflow And Reengineering
[13] MohanC.Recent trends in workflow management products,standards,and research. 1997.http://www.almaden.ibm.com/cs/exotica/wfnato97.ps
[14] Alonso G, Agrawal D,Abbadi E Aetal. Functionality and limitations of current workflow management systems. 1997. http://www.almaden.ibm.com/cs/exotica/wfmsys.ps
[15] E.Panagos, M. Rabinovich:Escalations in Workflow Management Systems, Procs. of DART'97, Rockville, Maryland, Nov.1997.
[16] J. Eder, E. Panagos, H. Pozewaunig, M. Rabinovich: Time Management in Workflow Systems. Proceedings of BIS'99. Poznan, Poland, 1999.
[17] S. Hwang, S. Ho, J. Tang. Mining Exception Instances to Facilitate Workflow Exception Handling. Proceedings of DASFAA'99, TaiWan, Apr. 1999.
[18] Hewlett-Packard. HP Changengine Technical Reference Guide. Edition 4.4. 2000
[19] McGregor.C. The Impact of Business Performance Monitoring on WfMC Standards. Presentation at the Workflow Management Coalition meeting. New York,NY. May 2001
[20] F. Casati, U. Dayal, D. Grigori, M.C. Shan. Improving Business Process Quality through Exception Understanding, Prediction, and Prevention. Procs. Of VLDB'01, Rome, Italy. Sept. 2001
[21] M. Berry, G. Linoff. Mastering Data Mining. Wiley, 2000.
[22] J.Hipp, U.Guntzer, and G. Nakaeizadeh. Algorithms for Association Rule Mining-A
General Survey and Comparison. In Proc.ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2000.
[23] R.Agrawal, T.Imielinski, and A.Swami. Mining association rules between sets of items in large databases. Processings of the ACM SIGMOD Conference on Management of data, pp.207-216, 1993.
[24] S.Brin, R.Motwani, J.D.Ullman, and S.Tsur. Dynamic Itemset counting and implication rules for market basket data. In ACM SIGMOD International Conference On the Management of Data. 1997.
[25] J.Kleinberg, C.Papadimitriou, and P.Raghavan. Segmentation problems. Proceedings of the 30th Annual Symposium on Theory of Computing, ACM. 1998.
[26] Vassiliadis. P., Quix.C., Vassiliou.Y. and Jarke,M. A Model for Data Warehouse Operational Processes. In Proceedings of the 12th International Conference on Advances Information Systems Engineering, eds. Bengt Wangler and Lars Bergman, Stockholm, 2000. pages 446-461
[27] J.Han, Y. Fu, W. Wang, K. Koperski, and O. Zaiane. DMQL:A data mining query language for relational databases. In SIGMOD'96 Workshop on Research Issues in Data Mining and Knowledge Discovery(DMKD'96), Montreal, Canada, 1996.
[28] Patterson.J.E. Workflow and Data Warehousing Taxonomy. Masters Thesis, University of Texas, Arlington. 1996
[29] List.B., Schiefer.J., Tjoa A.M. and Quirchmayr.G.T. The Process Warehouse-A Data Warehouse Approach for Business Process Management. In e-Business and Intelligent Management-Proceedings of the International Conference on Management of Information and Communication Technology. Copenhagen, Denmark. 1999
[30] Angela Bonifati, Fabio Casati, Umeshwar Dayal. Warehousing workflow data:Challenges and opportunities. In The VLDB Journal, pages 649-652, 2002
[31] McGregor.C. The Impact of Business Performance Monitoring on WfMC Standards. Presentation at the Workflow Management Coalition meeting. New York,NY. May 2001
[32] Kueng, P.. Process Performance Measurement System:a tool to support process-based organizations. In Total Quality Management, Vol. 11, No.1, January 2000. pages 67-86
[33] Vassiliadis. P., Quix.C., Vassiliou.Y. and Jarke,M. A Model for Data Warehouse Operational Processes. In Proceedings of the 12th International Conference on Advances Information Systems Engineering, eds. Bengt Wangler and Lars Bergman, Stockholm, 2000. pages 446-461
[34] 3.3 Decision Trees for Supervised Learning, http://qrb.mnsu.edu/grbts/doc/manual/J48 Decision Trees.html
[35] Decision Trees, http://www.cs.unimaas.nl/～postma/ml/AssignmentDT.doc
[36] http://www.microsoft.com/china/sql/70/gen/olap.htm
[37] http://java.sun.com/xml/jaxb
[2] Jiawei Han，Micheline Kamber著，《数据挖掘概念与理论》，机械工业出版社，2001．8
[3] Michael Corey等著，《Oracle8i数据仓库》，机械工业出版社，2002
[4] 飞思科技产品研发中心 编著，《Oracle9i数据仓库构建技术》，电子工业出版社，2003
[5] 谈竹奎 仇莉编著，《Oracle9i数据库管理员实用技术指南》，中国铁道出版社，2003
[6] Bjarki Holm，John Carnell等著，《Oracle9i Java程序设计》，清华大学出版社，2002
[7] Bruce Eckel著，《Thinking in Java》second Edition，机械工业出版社，2002
[8] 范玉顺主编，《工作流管理技术基础—实现企业业务过程重组、过程管理与过程自动化的核心技术》，清华大学出版社、施普林格出版社，2001
[9] 罗海滨、范玉顺、吴澄，工作流技术综述，《软件学报》，2000．11
[10] 李建强、范玉顺，基于Petri网的工作流模型性能分析方法，计算机应用，2001．6
[11] 肖晶、黄国兴等，基于流程日志的工作流性能分析和优化，《微型电脑应用》(已被录用)
[12] The Workflow Management Coalition (WfMC). http://www.wfmc.org/.Workflow And Reengineering
[13] MohanC.Recent trends in workflow management products,standards,and research. 1997.http://www.almaden.ibm.com/cs/exotica/wfnato97.ps
[14] Alonso G, Agrawal D,Abbadi E Aetal. Functionality and limitations of current workflow management systems. 1997. http://www.almaden.ibm.com/cs/exotica/wfmsys.ps
[15] E.Panagos, M. Rabinovich:Escalations in Workflow Management Systems, Procs. of DART'97, Rockville, Maryland, Nov.1997.
[16] J. Eder, E. Panagos, H. Pozewaunig, M. Rabinovich: Time Management in Workflow Systems. Proceedings of BIS'99. Poznan, Poland, 1999.
[17] S. Hwang, S. Ho, J. Tang. Mining Exception Instances to Facilitate Workflow Exception Handling. Proceedings of DASFAA'99, TaiWan, Apr. 1999.
[18] Hewlett-Packard. HP Changengine Technical Reference Guide. Edition 4.4. 2000
[19] McGregor.C. The Impact of Business Performance Monitoring on WfMC Standards. Presentation at the Workflow Management Coalition meeting. New York,NY. May 2001
[20] F. Casati, U. Dayal, D. Grigori, M.C. Shan. Improving Business Process Quality through Exception Understanding, Prediction, and Prevention. Procs. Of VLDB'01, Rome, Italy. Sept. 2001
[21] M. Berry, G. Linoff. Mastering Data Mining. Wiley, 2000.
[22] J.Hipp, U.Guntzer, and G. Nakaeizadeh. Algorithms for Association Rule Mining-A
General Survey and Comparison. In Proc.ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2000.
[23] R.Agrawal, T.Imielinski, and A.Swami. Mining association rules between sets of items in large databases. Processings of the ACM SIGMOD Conference on Management of data, pp.207-216, 1993.
[24] S.Brin, R.Motwani, J.D.Ullman, and S.Tsur. Dynamic Itemset counting and implication rules for market basket data. In ACM SIGMOD International Conference On the Management of Data. 1997.
[25] J.Kleinberg, C.Papadimitriou, and P.Raghavan. Segmentation problems. Proceedings of the 30th Annual Symposium on Theory of Computing, ACM. 1998.
[26] Vassiliadis. P., Quix.C., Vassiliou.Y. and Jarke,M. A Model for Data Warehouse Operational Processes. In Proceedings of the 12th International Conference on Advances Information Systems Engineering, eds. Bengt Wangler and Lars Bergman, Stockholm, 2000. pages 446-461
[27] J.Han, Y. Fu, W. Wang, K. Koperski, and O. Zaiane. DMQL:A data mining query language for relational databases. In SIGMOD'96 Workshop on Research Issues in Data Mining and Knowledge Discovery(DMKD'96), Montreal, Canada, 1996.
[28] Patterson.J.E. Workflow and Data Warehousing Taxonomy. Masters Thesis, University of Texas, Arlington. 1996
[29] List.B., Schiefer.J., Tjoa A.M. and Quirchmayr.G.T. The Process Warehouse-A Data Warehouse Approach for Business Process Management. In e-Business and Intelligent Management-Proceedings of the International Conference on Management of Information and Communication Technology. Copenhagen, Denmark. 1999
[30] Angela Bonifati, Fabio Casati, Umeshwar Dayal. Warehousing workflow data:Challenges and opportunities. In The VLDB Journal, pages 649-652, 2002
[31] McGregor.C. The Impact of Business Performance Monitoring on WfMC Standards. Presentation at the Workflow Management Coalition meeting. New York,NY. May 2001
[32] Kueng, P.. Process Performance Measurement System:a tool to support process-based organizations. In Total Quality Management, Vol. 11, No.1, January 2000. pages 67-86
[33] Vassiliadis. P., Quix.C., Vassiliou.Y. and Jarke,M. A Model for Data Warehouse Operational Processes. In Proceedings of the 12th International Conference on Advances Information Systems Engineering, eds. Bengt Wangler and Lars Bergman, Stockholm, 2000. pages 446-461
[34] 3.3 Decision Trees for Supervised Learning, http://qrb.mnsu.edu/grbts/doc/manual/J48 Decision Trees.html
[35] Decision Trees, http://www.cs.unimaas.nl/～postma/ml/AssignmentDT.doc
[36] http://www.microsoft.com/china/sql/70/gen/olap.htm
[37] http://java.sun.com/xml/jaxb