Interpretable Anomaly Prediction: Predicting anomalous behavior in industry 4.0 settings via regularized logistic regression tools
作者：
Highlights：
• We focus on the context of anomalous behavior prediction in Industry 4.0 settings.
• We provide a framework for supporting forthcoming faults forecasting.
• We provide a methodology for supporting predictive maintenance in decision making.
• We introduce a method called interpretable anomaly prediction.
• We build interpretable prediction models by using regularized logistic regression.
• The proposed technique predicts the probability that future data will be abnormal.
• Feature extraction and selection give insights on possible causes of failures.
• We provide an experimental evaluation related to a real-life chemical plant.
摘要
•We focus on the context of anomalous behavior prediction in Industry 4.0 settings.•We provide a framework for supporting forthcoming faults forecasting.•We provide a methodology for supporting predictive maintenance in decision making.•We introduce a method called interpretable anomaly prediction.•We build interpretable prediction models by using regularized logistic regression.•The proposed technique predicts the probability that future data will be abnormal.•Feature extraction and selection give insights on possible causes of failures.•We provide an experimental evaluation related to a real-life chemical plant.
论文关键词：
论文评审过程：Received 21 December 2018, Revised 11 July 2020, Accepted 17 August 2020, Available online 21 August 2020, Version of Record 30 November 2020.
论文官网地址：https://doi.org/10.1016/j.datak.2020.101850
