An Innovative Next Activity Prediction Approach Using Process Entropy and DAW-Transformer
Hadi Zare, Mostafa Abbasi, Maryam Ahang, Homayoun Najjaran
TL;DR
The paper tackles next-activity prediction in BPM by introducing an entropy-driven model selection framework and a novel DAW-Transformer that uses a dynamic window and full attribute integration to capture long-range dependencies. By quantifying process complexity with entropy, it assigns high-entropy datasets to the DAW-Transformer and low-entropy datasets to simpler, interpretable models, achieving strong accuracy on complex logs (e.g., Sepsis) while preserving interpretability on simpler logs (e.g., Road Traffic Fine). The methodology combines multi-feature embeddings, position encoding, and a transformer core, with explicit attention to process entropy in model choice. Results across six public datasets demonstrate clear regime-dependent performance gains, highlighting practical implications for adaptive, transparent BPM prediction in diverse operational settings.
Abstract
Purpose - In Business Process Management (BPM), accurate prediction of the next activities is vital for operational efficiency and decision-making. Current Artificial Intelligence (AI)/Machine Learning (ML) models struggle with the complexity and evolving nature of business process event logs, balancing accuracy and interpretability. This paper proposes an entropy-driven model selection approach and DAW-Transformer, which stands for Dynamic Attribute-Aware Transformer, to integrate all attributes with a dynamic window for better accuracy. Design/methodology/approach - This paper introduces a novel next-activity prediction approach that uses process entropy to assess the complexity of event logs and dynamically select the most suitable ML model. A new transformer-based architecture with multi-head attention and dynamic windowing mechanism, DAW-Transformer, is proposed to capture long-range dependencies and utilize all relevant event log attributes. Experiments were conducted on six public datasets, and the performance was evaluated with process entropy. Finding - The results demonstrate the effectiveness of the approach across these publicly available datasets. DAW-Transformer achieved superior performance, especially on high-entropy datasets such as Sepsis exceeding Limited window Multi-Transformers by 4.69% and a benchmark CNN-LSTM-SAtt model by 3.07%. For low-entropy datasets like Road Traffic Fine, simpler, more interpretable algorithms like Random Forest performed nearly as well as the more complex DAW-Transformer and offered better handling of imbalanced data and improved explainability. Originality/ value - This work's novelty lies in the proposed DAW-Transformer, with a dynamic window and considering all relevant attributes. Also, entropy-driven selection methods offer a robust, accurate, and interpretable solution for next-activity prediction.