New self-supervised framework enhances process prediction with data augmentation

Researchers have developed SiamSA-PPM, a novel self-supervised framework designed to improve predictive process monitoring (PPM) in scenarios with limited labeled data. This framework utilizes Siamese learning combined with statistically grounded data augmentation, employing unique transformation methods that consider control-flow semantics and common behavioral patterns. The approach generates realistic and semantically valid trace variants, enabling the learning of generalizable prefix representations without requiring labeled supervision. Experiments on real-world event logs demonstrate that SiamSA-PPM achieves competitive or superior performance in predicting both the next activity and the final outcome compared to existing state-of-the-art baselines. AI

IMPACT This research could improve the accuracy of predictive process monitoring in industries with scarce labeled data.

RANK_REASON The cluster describes a new academic paper detailing a novel research methodology. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• arXiv
• SiamSA-PPM
• Sjoerd Straten Van

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →