Control Chart Concurrent Pattern Classification Using Multi-Label Convolutional Neural Networks

Chuen-Sheng Cheng,Ying Ho,Pei-Wen Chen

doi:10.3390/app12020787

Chuen-Sheng Cheng, Ying Ho + Show 1 more

Open Access

https://doi.org/10.3390/app12020787

Copy DOI

Journal: Applied sciences	Publication Date: Jan 13, 2022
Citations: 5	License type: CC BY 4.0

Affiliation: Yuan Ze University, SHL Group (Taiwan)

Abstract

The detection and identification of non-random patterns is an important task in statistical process control (SPC). When a non-random pattern appears on a control chart, it means that there are assignable causes which will gradually deteriorate the process quality. In addition to the study of a single pattern, many researchers have also studied concurrent non-random patterns. Although concurrent patterns have multiple characteristics from different basic patterns, most studies have treated them as a special pattern and used the multi-class classifier to perform the classification work. This study proposed a new method that uses a multi-label convolutional neural network to construct a classifier for concurrent patterns of a control chart. This study used data from previous studies to evaluate the effectiveness of the proposed method with appropriate multi-label classification metrics. The results of the study show that the recognition performance of multi-label convolutional neural network is better than traditional machine learning algorithms. This study also used real-world data to demonstrate the applicability of the proposed method to online monitoring. This study aids in the further realization of smart SPC.

Highlights

Control charts are a statistical process control (SPC) tool used to determine if a manufacturing or business process is behaving as intended
It is clear that the proposed multi-label convolutional neural network (CNN) outperformed the feature-based SVC in classifying the concurrent pattern which is composed of cycle, upward trend and downward shift
The results demonstrate that CNN maintains high accuracy when basic constituents of concurrent patterns may not appear at the same point in time

Summary

Introduction

Control charts are a statistical process control (SPC) tool used to determine if a manufacturing or business process is behaving as intended. When a non-random pattern appears on the control chart, it suggests that there exist one or more assignable causes which will gradually deteriorate the process/product quality. To identify concurrent non-random patterns, researchers have often performed pre-processing to smooth the data or extract features. This study proposed a novel method of classifying single and concurrent non-random patterns on control charts through the use of a convolutional neural network (CNN) to build a multi-label classifier. To the best of our knowledge, this study is the first to apply a multi-label CNN classifier to control chart concurrent pattern classification. For CCPR, the translation invariance means that the classifier can recognize the pattern regardless of where it appears in the window This characteristic is extremely important when considering the dynamic nature of the non-random patterns [12]. The study is concluded and directions for future research are discussed

Methods

Multi-Label Convolutional Neural Network

Random Forest

Performance Metrics

Findings

Illustrative Example