A Novel Selection Algorithm of a Wavelet-Based Transformer Differential Current Features

Abstract

In this paper, a novel selection algorithm of wavelet-based transformer differential current features is proposed. The minimum description length with entropy criteria are employed for an initial selection of the mother wavelet and the resolution level, respectively; whereas stepwise regression is applied for obtaining the most statistically significant features. Dimensionality reduction is accordingly achieved, with an acceptable accuracy maintained for classification. The validity of the proposed algorithm is tested through a neuro-wavelet-based classifier of transformer inrush and internal fault differential currents. The proposed algorithm highlights the potential of utilizing synergism of integrating multiple feature selection techniques as opposed to an individual technique, which ensures optimal selection of the features.