一种能发现重复任务的过程挖掘算法

针对α-算法无法检测出重复任务的不足,对其进行了扩展,提出了一种机器学习和α-算法相结合的α*-算法。先建立任务的前驱/后继表,通过其中的信息学习出判定重复任务的启发式规则来标识日志中的所有重复任务,然后采用α-算法从标识后的日志中提取出工作流网,并对其进行调整得到包含重复任务的工作流网模型。实验结果说明了α*-算法的有效性。

Process mining algorithm to discover duplicate tasks

To overcome the disadvantage that α-algorithm is not able to detect the duplicate tasks,an improved algorithm called α~*-algorithm,was presented,which combines the techniques of machine learning and α-algorithm to distill duplicate tasks.In the algorithm,the Predecessor/Successor Tables of task are constructed first.Then the heuristic rules to determine the duplicate tasks are learned from the information contained in these tables and adopted to identify all duplicate tasks in the workflow log.Finally,the workflow net is extracted from the identified log using the α-algorithm and is fine-tuned to get the workflow model containing duplicate tasks.Experiment results illustrate the validity of the α~*-algorithm.