基于最优近似迹的增量日志过程挖掘优化方法

An Optimization Method for Mining Incremental Log Process Based on Optimal Approximation Trace

流程变更是业务流程管理的核心,增量日志是业务流程变更的核心,增量日志模型的挖掘在业务流程挖掘中显得尤为重要。目前已有的研究大部分是将原日志与增量日志合并挖掘,但原日志无法在该方法挖掘出的模型中重演;少部分是将增量日志当成噪音直接过滤,但增量日志对模型是有效且必须的。提出一种利用日志相似度来确定增量日志与原日志中迹的差异度,继而确定原模型的变化区域、修复原模型的方法。首先,根据给定的事件日志挖掘合理的过程模型;然后,通过迭代对比原日志与增量日志中迹的差异度,并在此基础上确定变化区域、基于修复规则修复模型;最后,通过实例分析验证了所提方法的可行性。

Business process change is the core of business process management,and incremental log is the core of business process change,thus incremental log model mining is particularly important in business process mining.Most of the existing researches combine the original log and incremental log mining,but the original log cannot be replayed in the model mined by this method.A small portion of researches is to filter incremental logs directly as noise,but incremental logs are valid and necessary for the model.In this paper,a method of using log similarity to determine the difference degree of trace between incremental log and original log,and then to determine the change area of original model to repair original model is proposed.Firstly,a reasonable process model is mined according to the given event log.Then,by iteratively comparing the trace difference between the original log and the incremental log,the variation area is determined on this basis,and the model is repaired based on the repair rules.Finally,an example is given to verify the feasibility of the proposed method.