Time Series Foundation Models for Process Model Forecasting
Yongbo Yu, Jari Peeperkorn, Johannes De Smedt, Jochen De Weerdt
TL;DR
This paper investigates Time Series Foundation Models (TSFMs) for Process Model Forecasting (PMF), reframing process evolution as forecasts over directly-follows (DF) time series derived from event logs. It compares zero-shot TSFM performance against baselines and explores parameter-efficient (LoRA) and full fine-tuning strategies across Chronos, MOIRAI, and TimesFM families. The results show zero-shot TSFMs typically outperform traditional methods, with adaptation offering modest gains that depend on dataset size and complexity. The findings highlight the data efficiency and cross-domain temporal structure transfer of TSFMs for PMF and call for broader evaluations and integration into process mining tools.
Abstract
Process Model Forecasting (PMF) aims to predict how the control-flow structure of a process evolves over time by modeling the temporal dynamics of directly-follows (DF) relations, complementing predictive process monitoring that focuses on single-case prefixes. Prior benchmarks show that machine learning and deep learning models provide only modest gains over statistical baselines, mainly due to the sparsity and heterogeneity of the DF time series. We investigate Time Series Foundation Models (TSFMs), large pre-trained models for generic time series, as an alternative for PMF. Using DF time series derived from real-life event logs, we compare zero-shot use of TSFMs, without additional training, with fine-tuned variants adapted on PMF-specific data. TSFMs generally achieve lower forecasting errors (MAE and RMSE) than traditional and specialized models trained from scratch on the same logs, indicating effective transfer of temporal structure from non-process domains. While fine-tuning can further improve accuracy, the gains are often small and may disappear on smaller or more complex datasets, so zero-shot use remains a strong default. Our study highlights the generalization capability and data efficiency of TSFMs for process-related time series and, to the best of our knowledge, provides the first systematic evaluation of temporal foundation models for PMF.