Matching business process behavior with encoding techniques via meta-learning: An anomaly detection study

Recording anomalous traces in business processes diminishes an event log?s quality. The abnormalities may represent bad execution, security issues, or deviant behavior. Focusing on mitigating this phenomenon, organizations spend efforts to detect their save resources and improve process execution. However, many real-world environments, reference models are unavailable, requiring expert assistance increasing costs. considerable number of techniques reduced availability experts pose additional challenge for particular scenarios. In work, we combine the representational power encoding with a Meta-learning strategy enhance detection logs towards fitting best discriminative capability between common irregular traces. Our approach creates log profile recommends most suitable technique increase anomaly performance. We used eight from different families, 80 descriptors, 168 logs, six types experiments. Results indicate that characteristics influence encodings. Moreover, investigate behavior?s choosing technique, demonstrating traditional mining analysis can be leveraged when matched intelligent decision support approaches.