Promoting Simple Agents: Ensemble Methods for Event-Log Prediction

2026-04-23

AI Summary

This work addresses the challenge of next-activity prediction in streaming event logs, where existing neural models such as LSTM and Transformer incur high computational overhead and exhibit unstable performance, while lightweight approaches like n-gram are efficient yet limited in accuracy. To bridge this gap, the authors propose a dynamic ensemble algorithm that adaptively selects the optimal predictor at runtime by concurrently activating only two models: an n-gram automaton and a context-window-optimized model. Experimental results on five real-world process mining datasets demonstrate that the proposed method significantly reduces memory consumption and inference latency while achieving prediction accuracy on par with or surpassing that of non-windowed neural baselines.

Technology Category

• Machine Learning: Ensemble Methods
• Natural Language Processing: Learning & Optimization for NLP
• Data Mining & Knowledge Management: Data Stream Mining

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Web Mining and Content Analysis: Models for Web evolution
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

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.

Problem

• event-log prediction
• next-activity prediction
• ensemble methods
• computational efficiency
• streaming event logs

Innovation

• ensemble methods
• n-grams
• event-log prediction
• promotion algorithm
• next-activity prediction