wavelet-based feature extraction and selection for classification of power system disturbances using support vector machines

this paper presents a new approach for the classification of the power system disturbances using support vector machines (svms). the proposed approach is carried out at three serial stages. firstly, the features to be form the svm classifier are obtained by using the wavelet transform and a few different feature extraction techniques. secondly, the features exposing the best classification accuracy of these features are selected by a feature selection technique called as sequential forward selection. thirdly, the best appropriate input vector for svm classifier is rummaged. the input vector is started with the first best feature and incrementally added the chosen features. after the addition of each feature, the performance of thesvmis evaluated. the kernel and penalty parameters of thesvmare determined by cross-validation. the parameter set that gives the smallest misclassification error is retained. finally, both the noisy and noiseless signals are applied to the classifier given above stages. experimental results indicate that the proposed classifier is robust and has more high classification accuracy with regard to the other approaches in the literature for this problem.