Leveraging Duration Pseudo-Embeddings in Multilevel LSTM and GCN Hypermodels for Outcome-Oriented PPM

2025-11-24

AI Summary

Existing predictive process monitoring (PPM) models struggle to capture irregular event durations and timestamp overlaps, limiting generalization across heterogeneous datasets. To address this, we propose a dual-input neural architecture that decouples event attributes from sequential temporal structure, and introduce a duration-aware pseudo-embedding matrix to explicitly encode dynamic temporal information. Furthermore, we design a self-tuning hypermodel that adaptively selects and fuses LSTM and GCN modules based on dataset-specific characteristics. Our approach significantly improves prediction accuracy and robustness across diverse balanced and imbalanced PPM tasks while reducing model complexity. Experimental results demonstrate that explicit temporal encoding enhances modeling fidelity in real-world PPM scenarios, improving both interpretability and structural learning capability.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Planning, Routing, and Scheduling: Temporal Planning
• Data Mining & Knowledge Management: Mining of Spatial, Temporal or Spatio-Temporal Data

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Existing deep learning models for Predictive Process Monitoring (PPM) struggle with temporal irregularities, particularly stochastic event durations and overlapping timestamps, limiting their adaptability across heterogeneous datasets. We propose a dual input neural network strategy that separates event and sequence attributes, using a duration-aware pseudo-embedding matrix to transform temporal importance into compact, learnable representations. This design is implemented across two baseline families: B-LSTM and B-GCN, and their duration-aware variants D-LSTM and D-GCN. All models incorporate self-tuned hypermodels for adaptive architecture selection. Experiments on balanced and imbalanced outcome prediction tasks show that duration pseudo-embedding inputs consistently improve generalization, reduce model complexity, and enhance interpretability. Our results demonstrate the benefits of explicit temporal encoding and provide a flexible design for robust, real-world PPM applications.

Problem

• Address temporal irregularities in predictive process monitoring
• Improve model adaptability across heterogeneous event datasets
• Enhance generalization and interpretability with duration encoding

Innovation

• Duration pseudo-embedding transforms temporal importance into learnable representations
• Dual input network separates event and sequence attributes for processing
• Self-tuned hypermodels enable adaptive architecture selection in models