One-sided control chart based on support vector machines with differential evolution algorithm

The statistical learning classification techniques have been successfully applied to statistical process control problems. In this paper, we proposed a one-sided control chart based on support vector machines (SVMs) and differential evolution (DE) algorithm to monitor a process with multivariate quality characteristics. The SVM classifier provides a continuous distance from the boundary, and the DE algorithm is used to obtain the optimal parameters of the SVM model by minimizing mean absolute error (MAE). The average run length of the proposed chart is computed using the Monte Carlo simulation approach. Several simulated cases are conducted using a multivariate normal distribution with 10 and 20 dimensions and three different process shift scenarios. In addition, we consider two non-normal distribution cases. The ARL performance of the proposed chart is better than the distance-based SVM chart. A real example is used to illustrate the application of the proposed control chart.