Detecting anomalies in production quality data using a method based on The Chi-Square Test Statistic

Authors Michael Mayr
Johannes Himmelbauer
Editors Min Song
Il-Yeol Song
Gabriele Kotsis
A Min Tjoa
Ismail Khalil
Title Detecting anomalies in production quality data using a method based on The Chi-Square Test Statistic
Booktitle DaWaK 2020: Big Data Analytics and Knowledge Discovery
Type in proceedings
Publisher Springer
Series Lecture Notes in Computer Science
Volume 12393
ISBN 978-3-030-59064-2
DOI 10.1007/978-3-030-59065-9_27
Month September
Year 2020
Pages 348-363
SCCH ID# 20063
Abstract

This paper describes the capability of the Chi-Square test statistic at detecting outliers in production-quality data. The goal is automated detection and evaluation of statistical anomalies for a large number of time series in the production-quality context. The investigated time series are the temporal course of sensor failure rates in relation to particular aspects (e.g. type of failure, information about products, the production process, or measuring sites). By means of an industrial use case, we show why in this setting our chosen approach is superior to standard methods for statistical outlier detection.