基于XGBoost增量实现业务流程执行结果的预测性监控方法

Incremental outcome-oriented predictive process monitoring based on XGBoost

随着工业制造业务流程智能化提升,以预测执行结果为目标的监控技术成为必需。该技术基于历史执行构建预测模型,从而对正在执行的流程进行结果预测。但现有研究假定流程执行行为一直保持不变,实际上流程在运行中常发生变化(即流程执行发生漂移),因此预测模型也需要适应这种漂移。针对这种情况,受到在线学习思想的启发,提出了基于XGBoost增量实现以流程执行结果为目标的预测流程监控技术,并分别在真实数据集和合成数据集上进行了大量的实验。实验结果表明,基于XGBoost的增量学习技术能够很好地为工业制造真实场景中的预测性流程监控提供一种有效的解决方案。

With the improvement of industrial manufacturing business processes,monitoring technology aimed at predicting the results of execution is necessary.The technique builds prediction models based on historical execution to predict the results of the processes being executed.However,existing studies assume that the process execution behavior remains the same,but the process often changes during the operation(the process execution drift)in practical application,so the prediction model needs to adapt to this drift.In response to this situation,inspired by the idea of online learning,a predictive process monitoring technology was proposed based on XGBoost incremental implementation targeting process execution outcomes,and a large number of experiments on real data sets and synthetic data sets were conducted respectively.The experimental results showed that the incremental learning technology based on XGBoost could well provide an effective solution for predictive process monitoring in real scenarios of industrial manufacturing.