Predictive Process Monitoring Using Object-centric Graph Embeddings

TL;DR

This paper introduces an end-to-end object-centric graph embedding model for predictive process monitoring, improving next activity and event time predictions by combining graph attention networks with LSTMs, evaluated on real and synthetic logs.

Contribution

It presents a novel model integrating graph attention and LSTM networks for object-centric process prediction tasks, advancing the state-of-the-art in predictive process monitoring.

Findings

Model achieves competitive performance on real and synthetic logs.

Graph attention networks effectively encode object relationships.

Combines temporal and relational information for improved predictions.

Abstract

Object-centric predictive process monitoring explores and utilizes object-centric event logs to enhance process predictions. The main challenge lies in extracting relevant information and building effective models. In this paper, we propose an end-to-end model that predicts future process behavior, focusing on two tasks: next activity prediction and next event time. The proposed model employs a graph attention network to encode activities and their relationships, combined with an LSTM network to handle temporal dependencies. Evaluated on one reallife and three synthetic event logs, the model demonstrates competitive performance compared to state-of-the-art methods.