Font Size: a A A

Study And Application On Support Vector Machine Classification

Posted on:2009-08-13Degree:MasterType:Thesis
Country:ChinaCandidate:Y J HuoFull Text:PDF
GTID:2178360242486916Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
Support Vector Machines (SVM) is a machine-learning algorithm based on statistical learning theory. Because of the excellent performance to limited samples, support vector machine is more and more widely used in fields such as pattern recognition,function fitting,fault diagnosis and so on. In this paper, we focused on the SVM classification problems, and such problems are analyzed especially. First, nonlinear classifiers algorithms of support vector machines and least squares support vector machines (LS-SVM) are discussed and compared. Then they are applied to data classification based on UCI data set. High accuracy is obtained. Furthermore on the basis of analyzing the parameter selection of Least Squares Support Vector Machine (LS-SVM) classifiers, a LS-SVM classification model is presented in which the parameters are optimized by particle swarm optimization(PSO) compared with Cross-Validation Method. Then they are applied to data classification based on UCI data set. Finally, PSO-(LS-SVM) classification algorithms are applied to power transformer fault diagnosis. The simulation results show that it meets the requirements of power transformer fault diagnosis for both convergence speed and calculation accuracy.
Keywords/Search Tags:Support Vector Machine, Least Squares Support Vector Machine, Particles Swarm optimization, Classifier, Power Transformer, Fault diagnosis
PDF Full Text Request
Related items