Promoting Simple Agents: Ensemble Methods for Event-Log Prediction
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866915952007839744 |
|---|---|
| author | Bollig, Benedikt Függer, Matthias Nowak, Thomas Zeinaty, Paul |
| author_facet | Bollig, Benedikt Függer, Matthias Nowak, Thomas Zeinaty, Paul |
| contents | We compare lightweight automata-based models (n-grams) with neural architectures (LSTM, Transformer) for next-activity prediction in streaming event logs. Experiments on synthetic patterns and five real-world process mining datasets show that n-grams with appropriate context windows achieve comparable accuracy to neural models while requiring substantially fewer resources. Unlike windowed neural architectures, which show unstable performance patterns, n-grams provide stable and consistent accuracy. While we demonstrate that classical ensemble methods like voting improve n-gram performance, they require running many agents in parallel during inference, increasing memory consumption and latency. We propose an ensemble method, the promotion algorithm, that dynamically selects between two active models during inference, reducing overhead compared to classical voting schemes. On real-world datasets, these ensembles match or exceed the accuracy of non-windowed neural models with lower computational cost. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_21629 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Promoting Simple Agents: Ensemble Methods for Event-Log Prediction Bollig, Benedikt Függer, Matthias Nowak, Thomas Zeinaty, Paul Machine Learning Artificial Intelligence Distributed, Parallel, and Cluster Computing Formal Languages and Automata Theory We compare lightweight automata-based models (n-grams) with neural architectures (LSTM, Transformer) for next-activity prediction in streaming event logs. Experiments on synthetic patterns and five real-world process mining datasets show that n-grams with appropriate context windows achieve comparable accuracy to neural models while requiring substantially fewer resources. Unlike windowed neural architectures, which show unstable performance patterns, n-grams provide stable and consistent accuracy. While we demonstrate that classical ensemble methods like voting improve n-gram performance, they require running many agents in parallel during inference, increasing memory consumption and latency. We propose an ensemble method, the promotion algorithm, that dynamically selects between two active models during inference, reducing overhead compared to classical voting schemes. On real-world datasets, these ensembles match or exceed the accuracy of non-windowed neural models with lower computational cost. |
| title | Promoting Simple Agents: Ensemble Methods for Event-Log Prediction |
| topic | Machine Learning Artificial Intelligence Distributed, Parallel, and Cluster Computing Formal Languages and Automata Theory |
| url | https://arxiv.org/abs/2604.21629 |