Support Vector Machines And Its Application In The Thermal Power Engineering Field

Statistical learning theory and support vector machine is new theory and technique,which can be used in the area of pattern recognition and be proved to be excellent inseveral application domains. Statistical learning theory and support vector machine isdescribed in detail, and the classification problems solved by support vector machine isstudied, at the same time, a practical validation in support vector regression is presented.The main works of this dissertation are summarized as follows: Statistical learningtheory and support vector machine are briefly introduced. Analysis of the most suitablestructures and optimum preferences of support vector machine for application to the faultdiagnosis is performed. On the basis of experiments results, much less amount ofsamples of support vector machine is studied, while some conclusions can be derivedfrom the application of support vector machine to diagnosis classification. Theapplication of support vector machine to regression is verified by practical examples,and according to the experiments, prediction results is reliable.