Abstract: Process mining aims to extract useful information from event logs,so as to discover,monitor and improve the actual business process.Most process mining technologies rely on standardized event logs,that is,each event in the event log corresponds to a case.However,the existing mining technology cannot deal with the unlabeled event log.To solve this problem,a case identification approach for unlabeled event log was proposed.Specifically,the degree of dependence between activities from the unlabeled event log was mined first according to the association rules,so as to mine the dependency between activities;according to the dependency relationship between activities,the possible activity relationships that was concurrency relation,exclusion relation and loop relation among activities was mined;finally a case identification algorithm was proposed to recognize the unlabeled event log and get the labeled event log.The proposed approach had been implemented in the open source platform ProM tool.Based on simulated log datasets and real log datasets,the effectiveness of this approach was verified.Through quantitative comparison with the current best approach in the field,the advantages of our approach were further verified.

Key words: process mining, Petri net, unlabeled event log, case identification

摘要： 流程挖掘旨在从事件日志中提取有用信息,从而发现、监控和改进实际的业务流程。大部分流程挖掘技术依赖于标准化的事件日志,即事件日志中的每个事件对应于一个案例。然而,已有的流程挖掘技术无法处理案例属性缺失的事件日志,即无标记事件日志。针对这个领域难题,提出一种面向无标记事件日志的案例识别方法。该方法首先根据关联规则从无标记事件日志中挖掘活动间的依赖程度,从而挖掘活动间的依赖关系;其次,根据活动间的依赖关系,挖掘活动间可能的活动关系,即并发关系、互斥关系和循环关系;最后提出一种案例识别算法对无标记事件日志进行案例识别,得到带有标记的事件日志。所提无标记事件日志案例识别方法已在开源平台ProM工具中实现。基于仿真日志数据集和真实日志数据集,验证了所提方法的有效性,通过与当前领域内最优方法进行定量比较,进一步验证了所提方法的优势。

关键词: 流程挖掘, Petri网, 无标记事件日志, 案例识别