A new weighted feature approach based on GA for speech recognition

A new weighted feature approach is shown how important to put weight factor on the feature vector of speech. Once the utterance comes with less discriminative, it would hard to capture the differences in the classification. Hence, the utterance divided into categories based on their influence in classification. This method is based on assumption that not all parts of utterance would appear balanced to provide good discriminative between them. By weighting parts, which most influence with higher value and vice versa, better distinguishing of utterances is possible. Using genetic algorithm, a new approach for weighting feature is introduced to improve recognition accuracy via exploitation of current recognition system simply by adding weight factor on feature vector.