Emotion Modeling in Speech Signals: Discrete Wavelet Transform and Machine Learning Tools for Emotion Recognition System

Table 3

The proposed method, formants, wavelet entropy, and the proposed method combined with formants and with entropy, along with the average accuracy in the last row.

Classifier type	Proposed method (%)	Formants (%)	Wavelet entropy (%)	Proposed method and formants (%)	Proposed method and entropy (%)
SVM	94.20	88.41	95.33	94.77	95.88
KNN	94.76	88.60	94.39	94.39	94.39
Efficient Logistic Regression	88.78	88.41	94.58	89.72	89.71
Naive Bayes	94.01	89.72	86.73	93.46	94.39
Ensemble	94.76	89.72	94.58	94.58	94.01
Neural Network	93.83	87.66	94.21	94.39	93.83
Average	93.39	88.75	93.30	93.55	93.70