Automatic voltage disturbance detection and classification using wavelets and multiclass logistic regression

This paper proposes new method for power quality disturbances classification using multiclass logistic regression. The features for the disturbances are extracted using wavelet packet transform and the rms value is used to characterize the magnitude of the disturbances. The detection and classification is done by employing machine learning. The proposed approach utilizes multiclass logistic regression with one against all strategy. The training and testing was done on seven different classes of simulated disturbances. The presented results show that the proposed method is able to produce classification with high-accuracy.