一种面向业务过程的剩余时间预测算法

A Residual Time Prediction Algorithm for Business Processes

准确的业务过程剩余时间预测有助于调整企业生产计划以及服务流程,以满足用户与市场的需求。现有方法多采用专家经验和历史数据构建的单一模型,具有无法智能选择较为符合当前业务过程模型的缺点。在传统变迁系统的基础上,提出一种基于多模型选择不同预测方法的算法。通过历史事件日志构建多种预测模型,基于模型数据纯度选择合适的预测模型对业务过程的剩余时间进行预测。通过在公开数据集上与传统变迁系统方法进行实验比较,验证了这种方法的有效性。

Accurate forecast of business process remaining time helps adjust the production plan and service flow of enterprises to meet the needs of users and the market.Most of the existing methods use a single model built with expert experience and historical data,which has the disadvantage of being unable to intelligently choose one which is more consistent with the current business process model.Based on the traditional transition system,this paper proposes an algorithm based on multiple models to select different prediction methods.A variety of prediction models are built through the historical event log,and the appropriate prediction model is selected based on the purity of model data to predict the remaining time of the business process.The validity of the proposed method is verified by comparing the open data set with the traditional transition system method.