Detecting human emotion via speech recognition by using ensemble classification model

Abstract

Speech Emotion Recognition is one of the most challenging researches in the field of Human-Computer Interaction (HCI). The accuracy of detecting emotion depends on several factors for example, type of emotion and number of emotion which is classified, quality of speech. In this research, we introduced the process of detecting 4 different emotion types (anger, happy, natural, and sad) from Thai speech which was recorded from Thai drama show which was most similar with daily life speech. The proposed algorithms used the combination of Support Vector Machine, Neural Network and k-Nearest Neighbors for emotion classification by using the ensemble classification method with majority weight voting. The experimental results show that emotion classification by using the ensemble classification method by using the majority weight voting can efficiency give the better accuracy results than the single model. The proposed method has better results when using with fundamental frequency (F0) and Mel-frequency cepstral coefficients (MFCC) of speech which give the accuracy results at 70.69%.