Building a recognition system of speech emotion and emotional states

To make a decision in companies or public organizations, the priority ordering plays an essential. For example, their discussion is essential for stakeholder to achieve mutual consensus,. In the discussion, the difference among consensus building processes can affect the last conclusion. Therefore, it is necessary for analysis to find critical remarks reaching the consensus ("focus remark"). However, it is a basis to confirm the gfocus remark" that the consensus building process can understand exactly from the disagreement state consent and detailed exposition parties. The consensus discussion is very helpful to promote interaction by the speech. The paper addresses the design of recognition system and results are achieved by means of MFCC (Mel Frequency Campestral Coefficients) and HMM (Hidden Markov Model). Results in recognition of six emotion patterns obtained 86.8% recognition rate. According to the relation of emotional states and emotions we analyzed the support more objectively.