Statistics for Speech recognition using Gaussian mixture model with MFCC feature extraction