<inline-formula><tex-math notation="LaTeX">${\sf PROPHET}$</tex-math><alternatives><mml:math><mml:mi mathvariant="sans-serif">PROPHET</mml:mi></mml:math><inline-graphic xlink:href="pasquadibisceglie-ieq1-3463487.gif"/></alternatives></inline-formula>: Explainable Predictive Process Monitoring With H (2024-11-01T00:00:00.000000Z)

Abstract

In this article, we introduce <inline-formula><tex-math notation="LaTeX">${\sf PROPHET}$</tex-math><alternatives><mml:math><mml:mi mathvariant="sans-serif">PROPHET</mml:mi></mml:math><inline-graphic xlink:href="pasquadibisceglie-ieq3-3463487.gif"/></alternatives></inline-formula>, an innovative approach to predictive process monitoring based on Heterogeneous Graph Neural Networks. <inline-formula><tex-math notation="LaTeX">${\sf PROPHET}$</tex-math><alternatives><mml:math><mml:mi mathvariant="sans-serif">PROPHET</mml:mi></mml:math><inline-graphic xlink:href="pasquadibisceglie-ieq4-3463487.gif"/></alternatives></inline-formula> is designed to strike a balance between accurate predictions and interpretability, particularly focusing on the next-activity prediction task. For this purpose, we represent the event traces recorded for different business process executions as heterogeneous graphs within a multi-view learning scheme combined with a heterogeneous graph learning approach. Using heterogeneous Graph Attention Networks (GATs), we achieve good accuracy by incorporating different characteristics of several events into graphs with different node types and leveraging different types of graph links to express relationships between event characteristics, as well as relationships between events. In addition, the use of a GAT model enables the integration of a modified version of the GNN Explainer algorithm to add the explainable component to the predictive model. In particular, the GNN Explainer algorithm is modified to disclose explainable information related to characteristics, events and relationships between events that mainly influenced the prediction. Experiments with various benchmark event logs prove the accuracy of <inline-formula><tex-math notation="LaTeX">${\sf PROPHET}$</tex-math><alternatives><mml:math><mml:mi mathvariant="sans-serif">PROPHET</mml:mi></mml:math><inline-graphic xlink:href="pasquadibisceglie-ieq5-3463487.gif"/></alternatives></inline-formula> compared to several current state-of-the-art methods and draw insights from explanations recovered through the GNN Explainer algorithm.

References36 items

How Universal Polynomial Bases Enhance Spectral Graph Neural Networks: Heterophily, Over-smoothing, and Over-squashing

Data- & compute-efficient deviance mining via active learning and fast ensembles

TSUNAMI - an explainable PPM approach for customer churn prediction in evolving retail data environments

A data-aware explainable deep learning approach for next activity prediction

HGExplainer: Explainable Heterogeneous Graph Neural Network

DARWIN : An online deep learning approach to handle concept drifts in predictive process monitoring

Multi-perspective enriched instance graphs for next activity prediction through graph neural network

Trace Encoding in Process Mining: a survey and benchmarking

Learnable Graph Convolutional Network and Feature Fusion for Multi-view Learning

HAQJSK: Hierarchical-Aligned Quantum Jensen-Shannon Kernels for Graph Classification

An Explainable Decision Support System for Predictive Process Analytics

Embedding Graph Convolutional Networks in Recurrent Neural Networks for Predictive Monitoring

FOX: a neuro-Fuzzy model for process Outcome prediction and eXplanation

Heterogeneous Graph Neural Network With Multi-View Representation Learning

How Attentive are Graph Attention Networks?

Process Outcome Prediction: CNN vs. LSTM (with Attention)

ProcessTransformer: Predictive Business Process Monitoring with Transformer Network

A Comparison of Deep-Learning Methods for Analysing and Predicting Business Processes

A Survey on Heterogeneous Graph Embedding: Methods, Techniques, Applications and Sources

Predictive Process Mining Meets Computer Vision

Learning Accurate LSTM Models of Business Processes

Predictive Business Process Monitoring with LSTM Neural Networks

Efficient Estimation of Word Representations in Vector Space

“JARVIS: Joiningadversarialtrainingwithvisiontransformersinnext-activitypre-diction,”

“Guest editorial: Deep neural networks for graphs: Theory, models, algorithms, and applications,”

Exploring Gated Graph Sequence Neural Networks for Predicting Next Process Activities

“GNNExplainer:Generatingexplanationsforgraphneuralnetworks,”in

A Comprehensive Survey on Graph Neural Networks

“Survey andcross-benchmarkcomparisonofremainingtimepredictionmethodsinbusinessprocessmonitoring,”

“Buildinginterpretablemodelsforbusinessprocessprediction usingsharedandspecialisedattentionmechanisms,”

data science from the University of Bari Aldo Moro, Italy. He is currently working toward the PhD degree withtheDepartmentofComputerScience,University ofBariAldoMoro,Italy

papers in international journals and conferences

“A direct data awareLSTMneuralnetworkarchitectureforcompleteremainingtraceandruntimeprediction,”

Donato Malerba (Senior Member, IEEE) professor with the Department of Computer University of Bari Aldo Moro, in Italy. He

“Explainabilityinprocessoutcomeprediction:Guidelinestoobtaininterpretableandfaithfulmodels,”

coordinator of Spoke 6 - Symbiotic AI within the project FAIR funded by NextGenerationEU