A Support Vector Machine for Recognizing Control Chart Patterns in Multivariate Processes

Statistical process control charts have been widely used for monitoring the quality characteristics of manufacturing processes. Analysis of unnatural patterns is an important aspect of control charting. The occurrence of unnatural patterns implies that a process is influenced by assignable causes, and corrective actions should be taken. In practice, there are many situations in which the simultaneous monitoring or control of two or more related quality characteristics is necessary. However, the traditional 2 T chart was insufficient for tasks of pattern recognition in multivariate processes. The main purpose of this paper is to develop a support vector machine-based (SVM) classifier for control chart pattern recognition in multivariate processes. In this study, we investigated four types of unnatural patterns identified in the literature, namely, trends, sudden shifts, mixtures, cyclic patterns. We considered multivariate processes with various covariance matrices to cover the whole range of parameters for each pattern type. The discriminant analysis in multivariate statistics is used as a baseline for comparison. A series of experiment s had been conducted to evaluate the sensitivity of the SVM-based classifier to the change of training parameters. The performance of the proposed approach was evaluated by computing its classification accuracy. Results from simulation studies indicate that the SVM-based classifier can perform significantly better than the discriminant analysis in identifying the multivariate unnatural patterns. We also proposed a two-stage classification procedure to further improve the performance of the SVM-based classifier. The proposed system may facilitate the diagnosis of out-of-control conditions in multivariate process control.