Faults diagnosis based on support vector machines and particle swarm optimization

Faults diagnosis is essentially one of pattern recognition problems. It has been gaining more and more attention to develop methods for improving the accuracy and effectiveness of pattern recognition. Support vector machine (SVM) is a powerful technique for the classification problems with small sampling, nonlinear and high dimension. However, one important problem encountered in setting up SVM models is how to determine the values of their parameters. The paper examined the diagnosis effects of SVMs with default and chosen parameters on the Steel Plates Faults Data Set, showing that different parameters may produce different diagnosis results. Particle swarm optimization (PSO), which is a heuristic method that optimizes a problem by iteratively trying to improve the candidate solution, was applied to optimize the parameters of SVMs, which enhanced the diagnosis accuracy.