Speech Emotion Pattern Recognition Agent in Mobile Communication Environment Using Fuzzy-SVM

In this paper, we propose a speech emotion recognition agent in mobile communication environment. The agent can recognize five emotional states - neutral, happiness, sadness, anger, and annoyance from the speech captured by a cellular-phone in real time. In general, the speech through the mobile network contains both speaker environmental noise and network noise, thus it can causes serious performance degradation due to the distortion in emotional features of the query speech. In order to minimize the effect of these noises and so improve the system performance, we adopt a simple MA (Moving Average) filter which has relatively simple structure and low computational complexity. Then a SFS (Sequential Forward Selection) feature optimization method is implemented to further improve and stabilize the system performance. For a practical application to call center problem, we created another emotional engine that distinguish two emotional states - ”agitation” which includes anger, happiness and annoyance, and ”calm” which includes neutral and sadness state. Two pattern classification methods, k-NN and Fuzzy-SVM, is compared for emotional state classifications. The experimental results indicate that the proposed method provides very stable and successful emotional classification performance as 72.5% over five emotional states and 86.5% over two emotional states.