Abstract: Fuzzy Cognitive Map（FCM） as a graph method analysis has application in the aspect of data classification. In order to improve the classification accuracy of speech emotion recognition and retain its advantage of the training speed, after a detailed analysis of feature level fusion and decision level fusion, fusion of FCM is proposed. To implement decision level fusion FCM, the numeric outputs of FCM is transformed to the probabilistic outputs, supplied a unified outputs for different input feature. On this basis, an adaptive weighted fusion method is introduced, which considers fully the difference accuracy rates of different features without the prior knowledge and subjective definition. Simulation experiments verify the correctness of the fusion FCM method. Compared to the traditional FCM and Support Vector Machine（SVM）, the proposed method has a higher recognition rate and greatly reduces the confusion level between the emotional.

Key words: Fuzzy Cognitive Map（FCM）, speech emotion recognition, decision level fusion, feature level fusion

摘要： 模糊认知图（Fuzzy Cognitive Map，FCM）作为一种图分析方法已在数据分类方面得到应用，为了提高其在语音情感识别中的分类精度，提出了融合FCM的方法。其中包括特征级融合和决策级融合两种方式。详细分析了这两种方式并提出将传统的模糊认知图的数值型输出转化为概率型输出，为不同特征提供了统一范围的初级识别结果。在此基础上，提出了自适应权值决策级融合方法。该方法充分考虑了分类器对不同特征的识别准确率差异。实验证明，提出的融合FCM方法相较于单一特征和单一分类器，具有更优的分类性能，同时大大降低了情感间的混淆程度。

关键词: 模糊认知图, 语音情感识别, 决策级融合, 特征级融合