中文摘要：

为了克服传统核判别方法将样本唯一归属于某一类的不足,基于模糊集理论对传统的核判别分析方法进行改进.引入了模糊类别隶属度并详细分析了基于模糊核判别分析的语音情感识别.识别实验中采用VQ方法进行了特征参数的融合,根据实验结果对模糊核判别分析和核判别分析的性能进行了比较和分析.

英文摘要：

Based on fuzzy class membership function,traditional kernel discriminant analysis （KDA） algorithm is improved.The speech emotion recognition on fuzzy kernel discriminant analysis algorithm is executed.In the experiment,firstly,the characteristic parameters are fused with VQ method.Then the performance of KDA and FKDA are compared on the results of experiment.