融合多维过程视角：一种基于上下文感知图注意力的业务流程预测框架

张希为; 方贤文; 毛古宝

doi:10.12263/DZXB.20250599

您当前的位置：

首页 >

文章列表页 >

融合多维过程视角：一种基于上下文感知图注意力的业务流程预测框架

学术论文 | 更新时间：2026-02-05

- 融合多维过程视角：一种基于上下文感知图注意力的业务流程预测框架
- Fusing Multi-Dimensional Process Views: A Context-Aware Graph Attention Framework for Business Process Prediction
- 电子学报 2025年53卷第10期页码：3705-3717
- 作者机构：
  
  1.安徽理工大学数学与大数据学院，安徽淮南 232001
  2.同济大学嵌入式系统与服务计算教育部重点实验室，上海 201804
- 作者简介：
  
  [ "张希为男，1999年4月出生于山东省泰安市.现为安徽理工大学数学与大数据学院博士研究生.主要研究方向为预测性流程监控、图神经网络等.E-mail: zxw@aust.edu.cn" ]
  [ "方贤文男，1975年10月出生于河南省信阳市.现为安徽理工大学数学与大数据学院教授、博士生导师.主要研究方向为流程挖掘、Petri网等.E-mail: xwfang@aust.edu.cn" ]
  [ "毛古宝男，1997年7月出生于安徽省安庆市.现为安徽理工大学数学与大数据博士研究生.主要研究方向为预测性流程监控、异常检测等.E-mail: mgb@aust.edu.cn" ]
- 基金信息：
  
  国家自然科学基金(61572035;61402011);嵌入式系统与服务计算教育部重点实验室开放课题(ESSCKF2021-05);安徽省重点研究与开发计划项目(2022a05020005);安徽省自然科学基金（水科学联合基金）(2308085US11);安徽省高校领军骨干人才项目(2020-01-12);安徽省学术和技术带头人资助项目(2022D327)
- DOI：10.12263/DZXB.20250599
  中图分类号： TP391;
- 收稿：2025-07-09，
  
  录用：2025-10-09，
  
  纸质出版：2025-10-25
- 稿件说明：
移动端阅览
张希为, 方贤文, 毛古宝. 融合多维过程视角：一种基于上下文感知图注意力的业务流程预测框架[J]. 电子学报, 2025, 53(10): 3705-3717.

ZHANG Xi-wei, FANG Xian-wen, MAO Gu-bao. Fusing Multi-Dimensional Process Views: A Context-Aware Graph Attention Framework for Business Process Prediction[J]. Acta Electronica Sinica, 2025, 53(10): 3705-3717.
张希为, 方贤文, 毛古宝. 融合多维过程视角：一种基于上下文感知图注意力的业务流程预测框架[J]. 电子学报, 2025, 53(10): 3705-3717. DOI：10.12263/DZXB.20250599

ZHANG Xi-wei, FANG Xian-wen, MAO Gu-bao. Fusing Multi-Dimensional Process Views: A Context-Aware Graph Attention Framework for Business Process Prediction[J]. Acta Electronica Sinica, 2025, 53(10): 3705-3717. DOI：10.12263/DZXB.20250599

摘要

随着数字化转型的深化，以预测性流程监控为核心的数据驱动流程分析技术，已成为企业提升运营效能与决策水平的关键.为提升预测性流程监控的精度与泛化能力，现有研究致力于从海量的历史事件日志中挖掘流程的深层表征.然而，真实业务流程的演化不仅遵循既定的时序逻辑，也受到资源分配、数据依赖等潜在结构化因素的影响，这对现有预测模型的表征能力构成了严峻挑战.具体而言，主流预测方法的性能常受限于其单一的过程视角与静态的信息融合策略.多数方法，即便是基于图神经网络，也倾向于从单一的控制流视角建模，忽略了资源交互、数据依赖等关键维度，进而造成流程深层结构与多维关系表征的鸿沟.此外，少数尝试融合多维信息的研究也大多采用静态融合策略，缺乏对多维信息的上下文感知融合能力，使得模型适应性不足.为应对上述挑战，本文提出一种上下文感知多视角图融合预测框架（Context-Aware Multi-view Graph Fusion，CAM-GF）.该框架首先突破控制流局限，系统性地构建了一个过程图谱，该图谱不仅包含捕捉宏观规律的长期依赖图等基础控制流视角，还涵盖了如揭示组织协作关系的资源交互图等扩展语义视角，以捕获流程全局性、多层次的结构化知识.进而，在时空信息融合层面，设计了一种新颖的上下文感知图注意力机制，它以案例的实时执行前缀为输入，动态学习并分配各视角的融合权重.最后，引入Transformer对动态融合后的特征序列进行深度时序建模，以实现对下一活动的精准预测.为验证框架的有效性与实用价值，本文在6六个公开的真实业务流程数据集上进行了综合实验.结果表明，相较于多种主流模型，CAM-GF框架在下一活动预测任务上预测准确率平均提升4.16个百分点.此外，框架动态生成的注意力权重为模型行为提供了高价值的可解释性，揭示了模型如何根据预测反馈与实时上下文，既能在局部上下文失效时回归对全局流程结构的依赖，也能在特定情境下转而聚焦于资源分配等关键语义视角，充分验证了所提框架在精度与透明度上的先进性与实用价值.

Abstract

Amidst deepening digital transformation

data-driven process analysis

with Predictive Process Monitoring (PPM) at its core

has become pivotal for enhancing enterprise operational efficiency and decision-making. To improve the accuracy and generalization of PPM

existing research focuses on mining deep representations from vast event logs. However

the evolution of real-world business processes is influenced not only by temporal logic but also by underlying structural factors such as resource allocation and data dependencies. This complexity poses a formidable challenge to the representational capabilities of existing predictive models. Specifically

the performance of mainstream predictive methods is often constrained by their reliance on a singular process view and static information fusion strategies. Most approaches

even those based on Graph Neural Networks (GNNs)

tend to model processes from a single control-flow perspective. This overlooks critical dimensions such as resource interactions and data dependencies

creating a gap in the representation of deep process structures and multi-dimensional relationships. Furthermore

the few studies that attempt to integrate multi-dimensional information typically employ static fusion strategies

lacking a context-aware fusion capability and resulting in models with insufficient adaptability. To address these challenges

this paper proposes a context-aware multi-view graph fusion (CAM-GF) framework. The framework first transcends the limitations of the control-flow perspective by systematically constructing a process graph map. This graph map comprises not only basic control-flow views

such as a long-term dependency graph that captures macroscopic patterns

but also extended semantic views

like a resource interaction graph that reveals organizational collaboration

thereby capturing holistic and multi-level structural knowledge. Subsequently

a novel context-aware graph attention mechanism is designed for spatio-temporal information fusion. It takes the real-time prefix of a case as input to dynamically learn and assign fusion weights to each view. Finally

a Transformer is introduced to perform deep temporal modeling on the dynamically fused feature sequence to achieve precise next-activity prediction. To validate the framework’s effectiveness and practical value

comprehensive experiments were conducted on six public

real-world business process datasets. The results demonstrate that

compared to various mainstream baseline models

the CAM-GF framework achieves an average accuracy improvement of 4.16 percentage points on the next-activity prediction task. Furthermore

the dynamically generated attention weights provide high-value interpretability for the model’s behavior

revealing how the model

based on predictive feedback and real-time context

can both rely on global process structures when local context fails and pivot to focus on critical semantic views

such as resource allocation

in specific situations. This thoroughly validates the proposed framework’s advancement in both accuracy and transparency.

关键词

Keywords

references

NEU D A , LAHANN J , FETTKE P . A systematic literature review on state-of-the-art deep learning methods for process prediction [J ] . Artificial Intelligence Review , 2022 , 55 ( 2 ): 801 - 827 .

RAMA-MANEIRO E , VIDAL J C , LAMA M . Deep learning for predictive business process monitoring: Review and benchmark [J ] . IEEE Transactions on Services Computing , 2023 , 16 ( 1 ): 739 - 756 .

TAX N , VERENICH I , LA ROSA M , et al . Predictive business process monitoring with LSTM neural networks [M ] // Advanced Information Systems Engineering . Cham : Springer , 2017 : 477 - 492 .

CAMARGO M , DUMAS M , GONZÁLEZ-ROJAS O . Learning accurate LSTM models of business processes [M ] // Business Process Management . Cham : Springer , 2019 : 286 - 302 .

DI FRANCESCOMARINO C , GHIDINI C , MAGGI F M , et al . An eye into the future: Leveraging a-priori knowledge in predictive business process monitoring [M ] // Business Process Management . Cham : Springer International Publishing , 2017 : 252 - 268 .

徐兴荣 , 刘聪 , 李婷 , 等 . 基于双向准循环神经网络和注意力机制的业务流程剩余时间预测方法 [J ] . 电子学报 , 2022 , 50 ( 8 ): 1975 - 1984 .

XU X R , LIU C , LI T , et al . Business process remaining time prediction: An approach based on bidirectional quasi recurrent neural network with attention [J ] . Acta Electronica Sinica , 2022 , 50 ( 8 ): 1975 - 1984 . (in Chinese)

WICKRAMANAYAKE B , HE Z P , OUYANG C , et al . Building interpretable models for business process prediction using shared and specialised attention mechanisms [J ] . Knowledge-Based Systems , 2022 , 248 : 108773 .

RAMA-MANEIRO E , VIDAL J C , LAMA M . Embedding graph convolutional networks in recurrent neural networks for predictive monitoring [J ] . IEEE Transactions on Knowledge and Data Engineering , 2024 , 36 ( 1 ): 137 - 151 .

WU Z H , PAN S R , CHEN F W , et al . A comprehensive survey on graph neural networks [J ] . IEEE Transactions on Neural Networks and Learning Systems , 2021 , 32 ( 1 ): 4 - 24 .

万升 , 杨健 , 宫辰 . 基于图神经网络的高光谱图像分类研究进展 [J ] . 电子学报 , 2023 , 51 ( 6 ): 1687 - 1709 .

WAN S , YANG J , GONG C . Advances of hyperspectral image classification based on graph neural networks [J ] . Acta Electronica Sinica , 2023 , 51 ( 6 ): 1687 - 1709 . (in Chinese)

VENUGOPAL I , TÖLLICH J , FAIRBANK M , et al . A comparison of deep-learning methods for analysing and predicting business processes [C ] // 2021 International Joint Conference on Neural Networks . Piscataway : IEEE , 2021 : 1 - 8 .

WEINZIERL S . Exploring gated graph sequence neural networks for predicting next process activities [M ] // Business Process Management Workshops . Cham : Springer International Publishing , 2022 : 30 - 42 .

AMIRI ELYASI K , VAN DER AA H , STUCKENSCHMIDT H . PGTNet: A process graph Transformer network for remaining time prediction of business process instances [M ] // Advanced Information Systems Engineering . Cham : Springer Nature Switzerland , 2024 : 124 - 140 .

BUKHSH Z A , SAEED A , DIJKMAN R M . ProcessTransformer: Predictive business process monitoring with transformer network [EB/OL ] . ( 2021-04-01 )[ 2025-06-30 ] . https://arXiv.org/abs/2104.00721 https://arXiv.org/abs/2104.00721 .

DE SMEDT J , DE WEERDT J . Predictive process model monitoring using long short-term memory networks [J ] . Engineering Applications of Artificial Intelligence , 2024 , 133 : 108295 .

TAYMOURI F , LA ROSA M , ERFANI S , et al . Predictive business process monitoring via generative adversarial nets: The case of next event prediction [EB/OL ] .( 2020-04-01 )[ 2025-06-30 ] . https://arXiv.org/abs/2003.11268 https://arXiv.org/abs/2003.11268 .

EVERMANN J , REHSE J R , FETTKE P . Predicting process behaviour using deep learning [J ] . Decision Support Systems , 2017 , 100 : 129 - 140 .

CHIORRINI A , DIAMANTINI C , GENGA L , et al . Multi-perspective enriched instance graphs for next activity prediction through graph neural network [J ] . Journal of Intelligent Information Systems , 2023 , 61 ( 1 ): 5 - 25 .

WANG J X , LU C L , YU Y F , et al . HiGPP: A history-informed graph-based process predictor for next activity [M ] // Service-Oriented Computing . Singapore : Springer Nature Singapore , 2024 : 337 - 353 .

PASQUADIBISCEGLIE V , SCARINGI R , APPICE A , et al . PROPHET: Explainable predictive process monitoring with heterogeneous graph neural networks [J ] . IEEE Transactions on Services Computing , 2024 , 17 ( 6 ): 4111 - 4124 .

李鑫 , 陆伟 , 马召祎 , 等 . 基于图注意力和改进Transformer的节点分类方法 [J ] . 电子学报 , 2024 , 52 ( 8 ): 2799 - 2810 .

LI X , LU W , MA Z Y , et al . A node classification method based on graph attention and improved Transformer [J ] . Acta Electronica Sinica , 2024 , 52 ( 8 ): 2799 - 2810 . (in Chinese)

VAN DER AALST W . Data science in action [M ] // Process Mining . Berlin, Heidelberg : Springer , 2016 : 3 - 23 .

VASWANI A , SHAZEER N , PARMAR N , et al . Attention is all you need [C ] // Proceedings of the 31st International Conference on Neural Information Processing Systems . Red Hook : Curran Associates Inc. , 2017 : 6000 - 6010 .

SINDHGATTA R , MOREIRA C , OUYANG C , et al . Exploring interpretable predictive models for business processes [M ] // Business Process Management . Cham : Springer , 2020 : 257 - 272 .

DONG L L , LIU C , REN C G . GCN-ONLSTM: Process next event prediction method based on spatio-temporal feature fusion [J ] . International Journal of Science and Engineering Applications , 2023 , 12 ( 3 ): 1 - 4 .

WANG J X , YU Y F , FANG N , et al . MHG-predictor: A multi-layer heterogeneous graph-based predictor for next activity in complex business processes [C ] // Proceedings of the ACM on Web Conference 2025 . New York : ACM , 2025 : 500 - 509 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

面向无蜂窝大规模MIMO的分布式智能内容缓存与用户关联联合优化方法

从博弈论视角解构去噪扩散概率模型的视觉概念生成机制

Na₅Eu(MO₄)₄(M=Mo,W)的B(λkq)强度参数和跃迁几率

Ho³⁺和Er³⁺在YLiF₄晶体中的光谱强度和Ω_λ参数

1.3μmInGaAsP/InP发光管退化的研究