基于Stacking策略的过程剩余执行时间预测

Process Remaining Execution Time Prediction Based on Stacking Strategy

为企业提供精确高效的业务过程剩余执行时间预测对企业合理安排后续计划有重要意义.为了体现各个基本活动以及活动之间存在的依赖关系对整体结果的影响,本文结合朴素贝叶斯、支持向量回归与长短期记忆网络提出一种基于Stacking策略进行模型融合的业务过程剩余时间预测算法.为了验证模型,将该方法应用于2017年荷兰银行贷款申请真实数据集中.实验结果表明,我们的方法与相关的业务过程时间剩余预测方法相比具有更高的预测精度以及更好的预测效果.

Providing accurate and efficient business process remaining execution time prediction is of great significance for enterprises to arrange follow-up plans reasonably.In order to express the influence of various basic activities and the dependencies between activities on the overall results,this paper proposes a business process remaining time prediction algorithm based on Stacking strategy with naive bayes classifier,support vector regression and long short-term memory networks.The method is applied to the real data set of the2017 Dutch bank loan application.The experimental results show that our method has higher prediction accuracy and better prediction effect than the related business process time remaining prediction method.