Home // DATA ANALYTICS 2021, The Tenth International Conference on Data Analytics // View article

Detection of Concept Drift in Manufacturing Data with SHAP Values to Improve Error Prediction

Authors:
Christian Seiffer
Holger Ziekow
Ulf Schreier
Alexander Gerling

Keywords: AI in manufacturing, error prediction, concept drift detection, clustering, SHAP values

Abstract:
Production processes are inherently subject to dynamic change. This makes the extraction of error causes and the prediction of errors from manufacturing data by machine learning (ML) a difficult challenge, but at the same time it is the key to improve product quality and thus to economic profit. As part of the PREFERML research project (Proactive Error Avoidance in Production through Machine Learning), we present a method for detecting concept drift using clustering based on SHAP values (SHapley Additive exPlanations) and propose strategies to handle concept drift. Evaluation based on real manufacturing data shows that the cluster specific approach improves concept drift detection and can yield economic benefits.

Pages: 51 to 60

Copyright: Copyright (c) IARIA, 2021

Publication date: October 3, 2021

Published in: conference

ISSN: 2308-4464

ISBN: 978-1-61208-891-4

Location: Barcelona, Spain

Dates: from October 3, 2021 to October 7, 2021