Feature selection using Fisher's ratio technique for automatic speech recognition

TL;DR

This paper explores using Fisher's ratio to select the most discriminative MFCC features for automatic speech recognition, aiming to improve classification accuracy with fewer features.

Contribution

It introduces a Fisher's ratio-based feature selection method for MFCCs in ASR, enhancing classification efficiency and accuracy.

Findings

Selected MFCC coefficients improve classification accuracy.

Fisher's ratio effectively identifies discriminative features.

Reduced feature set maintains or enhances recognition performance.

Abstract

Automatic Speech Recognition involves mainly two steps; feature extraction and classification . Mel Frequency Cepstral Coefficient is used as one of the prominent feature extraction techniques in ASR. Usually, the set of all 12 MFCC coefficients is used as the feature vector in the classification step. But the question is whether the same or improved classification accuracy can be achieved by using a subset of 12 MFCC as feature vector. In this paper, Fisher's ratio technique is used for selecting a subset of 12 MFCC coefficients that contribute more in discriminating a pattern. The selected coefficients are used in classification with Hidden Markov Model algorithm. The classification accuracies that we get by using 12 coefficients and by using the selected coefficients are compared.