Enhancing speech recognition using minimal weighted mel frequency cepstral coefficients and improved particle swarm optimization
