DESCRIPTIVE STATISTICS METHOD IN ISOLATED MALAY DIGITS FEATURE EXTRACTION

Nowadays, the use of speech recognition feature extraction methods are not optimal in terms of accuracy and speed when they are applied to a specific environment and recognition task. The performance of the speech recognition system depends on the feature extraction stage and classification stage. In this paper, descriptive statistics were used after feature extraction stage to minimize the amount of feature vector elements and to maximize the peak amplitude in isolated Malay Digit speech recognition. Artificial Neural Network (ANN) was used as classifier to evaluate these new feature vectors' representations. The obtained speech recognition rate was 96.67 %. Therefore, this method shows an improvement in the recognition rate.