Figure 10From: Evaluation of influence of spectral and prosodic features on GMM classification of Czech and Slovak emotional speech Influence of the type of the feature vector on GMM emotion classification error rate; N gmix = 6, N iter = 1200. Back to article page