计算机科学与探索 ›› 2015, Vol. 9 ›› Issue (12): 1523-1530.DOI: 10.3778/j.issn.1673-9418.1503037

• 人工智能与模式识别 • 上一篇    下一篇

高速公路冰雪应急流程挖掘与模型改进方法

孙亚男+,杜玉越,刘  伟   

  1. 山东科技大学 信息科学与工程学院,山东 青岛 266590
  • 出版日期:2015-12-01 发布日期:2015-12-04

Highway Snow and Ice Emergency Response Process Mining and Model Improvement

SUN Yanan+, DU Yuyue, LIU Wei   

  1. College of Information Science and Engineering, Shandong University of Science and Technology, Qingdao, Shandong 266590, China
  • Online:2015-12-01 Published:2015-12-04

摘要: 为了优化和管理业务流程,需要从大量的事件日志中挖掘模型,并对模型进行改进和检测。基于高速公路冰雪事件应急响应流程执行事件日志,在ProM平台上,通过对多种挖掘算法的对比实验,最终选择运用α算法对事件日志进行业务流程挖掘,获得其工作流网模型。由于α算法在处理某些特殊流程结构和情况时,如不可见变迁、重命名任务、特殊循环结构和非自由选择结构等,存在挖掘结果不准确问题,进而根据实际业务规则,对挖掘的模型进行改进。最后通过一致性检测验证了改进模型拟合度较高。

关键词: 业务流程;工作流网;模型改进;一致性检测;&alpha, 算法

Abstract: In order to optimize and manage the business processes, it needs to discover the model from a large number of event logs and improve and check the model. Based on the highway snow and ice emergency response process event logs, this paper makes comparison experiments among different algorithms on the ProM platform, finally chooses α-algorithm to mine the business processes from event logs and obtains the workflow net model. The α-algorithm deals with some special process structures and situations inaccurately, for example, invisible tasks, rename tasks, special loop structure and non-free choice process constructs. So this paper improves the mined model according to the realistic business rule and checks the improved model. Finally this paper proves that the improved model has a perfect fitness through conformance checking.

Key words: business process, workflow net, model improvement, conformance checking; α-algorithm