Abstract :
Features which are linearly dependent on other features do not contribute toward pattern classification by linear techniques. In order to detect the linearly dependent features, a measure of linear dependence is proposed. This measure is used as an aid in feature selection. Speaker verification experiments demonstrate the usefulness of employing such a measure in practice.
Keywords :
Analysis of variance, dependent features, feature selection, pattern classification, speaker verification.; Distortion measurement; Encoding; Euclidean distance; Mutual information; Pattern recognition; Performance evaluation; Probability; Rate distortion theory; Signal to noise ratio; Testing; Analysis of variance, dependent features, feature selection, pattern classification, speaker verification.;
