Statistical process control charts have been widely utilized in manufacturing processes for determining whether a process is run in its intended mode or in the presence of unnatural patterns, it's a multi-class classifier problem. Effective approaches to recognize control chart patterns is essential for a manufacturing process to maintain high-quality products. This paper we use the Directed Acyclic Graph (DAG) tree learning architecture, which combines many two-class classifiers together to solve the multi-class classifier problem. For each node we chose the support vector machine (SVM) using a particle swarm optimization (PSO) algorithm to optimize the parameter of the SVM kernel function. Here the PSO not only takes the kernel function parameters as variables but also the feature vector of the SVM to optimize. Simulation results show the propose algorithm achieves a high recognition accuracy and solve the unable recognition area.
展开▼
