Promoting Simple Agents: Ensemble Methods for Event-Log Prediction
Benedikt Bollig, Matthias Függer, Thomas Nowak, Paul Zeinaty ✉ · 2026
Read Promoting Simple Agents: Ensemble Methods for Event-Log Prediction by Benedikt Bollig, Matthias Függer, Thomas Nowak, Paul Zeinaty ✉ (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-21629.txt.
Description
Benedikt Bollig and three colleagues compare light n-gram models with LSTM and Transformer models for next-activity prediction in streams. On synthetic patterns and five real sets, n-grams with suitable windows match neural accuracy, needing far fewer resources. Unlike windowed models, n-grams hold. Voting lifts n-gram scores but needs many parallel agents, raising memory and latency. They propose the promotion algorithm, picking two models in inference, cutting overhead versus classical voting. On real sets, these ensembles match or beat non-windowed accuracy for less.
Catalog record
| CREATOR | Benedikt Bollig, Matthias Függer, Thomas Nowak, Paul Zeinaty ✉ |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 36957 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.21629