Machine Learning in Business Process Monitoring: A Comparison of Deep Learning and Classical Approaches Used for Outcome Prediction

TLDR

Predictive process monitoring forecasts business process behavior, performance, and outcomes at runtime, enabling early problem detection and resource reallocation, yet most outcome-oriented approaches rely on classical machine learning despite deep learning’s recent breakthroughs and limited understanding of which event log properties favor DL. The study compares the performance of deep learning models (feedforward neural networks and LSTMs) with classical machine learning techniques (random forests and SVMs) on five publicly available event logs. The authors evaluated these models using five public event logs, assessing predictive accuracy across different log characteristics. Deep learning models consistently outperform classical machine learning, especially on logs with high variant‑to‑instance ratios, perform more stably with imbalanced targets and high event‑to‑activity ratios, and LSTMs excel when input data is predominantly generated at runtime, with these results generalizing beyond the sampled logs.

Abstract

Abstract Predictive process monitoring aims at forecasting the behavior, performance, and outcomes of business processes at runtime. It helps identify problems before they occur and re-allocate resources before they are wasted. Although deep learning (DL) has yielded breakthroughs, most existing approaches build on classical machine learning (ML) techniques, particularly when it comes to outcome-oriented predictive process monitoring. This circumstance reflects a lack of understanding about which event log properties facilitate the use of DL techniques. To address this gap, the authors compared the performance of DL (i.e., simple feedforward deep neural networks and long short term memory networks) and ML techniques (i.e., random forests and support vector machines) based on five publicly available event logs. It could be observed that DL generally outperforms classical ML techniques. Moreover, three specific propositions could be inferred from further observations: First, the outperformance of DL techniques is particularly strong for logs with a high variant-to-instance ratio (i.e., many non-standard cases). Second, DL techniques perform more stably in case of imbalanced target variables, especially for logs with a high event-to-activity ratio (i.e., many loops in the control flow). Third, logs with a high activity-to-instance payload ratio (i.e., input data is predominantly generated at runtime) call for the application of long short term memory networks. Due to the purposive sampling of event logs and techniques, these findings also hold for logs outside this study.