Log Data Preparation for Predicting Critical Errors Occurrences

Myriam Lopez (1) , Marie Beurton-Aimar (1) , Gayo Diallo (1, 2) , Sofian Maabout (1)

1 LaBRI - Laboratoire Bordelais de Recherche en Informatique
2 BPH - Bordeaux population health

Failure anticipation is one of the key industrial research objectives with the advent of Industry 4.0. This paper presents an approach to predict high importance errors using log data emitted by machine tools. It uses the concept of bag to summarize events provided by remote machines, available within log files. The idea of bag is inspired by the Multiple Instance Learning paradigm. However, our proposal follows a different strategy to label bags, that we wanted as simple as possible. Three main setting parameters are defined to build the training set allowing the model to fine-tune the trade-off between early warning, historic informativeness and forecast accuracy. The effectiveness of the approach is demonstrated using a real industrial application where critical errors can be predicted up to seven days in advance thanks to a classification model.

Mots clés: Predictive Maintenance (PdM); Machine Learning (ML); Classification; Log data; data preparation

Domaines: Informatique [cs]; Machine Learning [stat.ML]; Apprentissage [cs.LG]; Réseau de neurones [cs.NE]; Analyse numérique [cs.NA]

Myriam Lopez, Marie Beurton-Aimar, Gayo Diallo, Sofian Maabout. Log Data Preparation for Predicting Critical Errors Occurrences. Trends and Applications in Information Systems and Technologies, 1366, pp.224-233, 2021, Advances in Intelligent Systems and Computing, ⟨10.1007/978-3-030-72651-5_22⟩. ⟨hal-03195735⟩