摘要
传统的K近邻算法存在误判风险,针对其不足提出了一种基于模糊K近邻的语音情感识别算法,通过引入模糊隶属度的概念,求出不同的特征参数对于不同情感识别的贡献度,并将其与欧式距离加权应用于语音情感识别中,实验验证了算法的有效性.
To solve the limitation that the traditional K-nearest neighbor algorithm may divide a sample to a wrong class,the concept of fuzzy class membership function is used to improve the K-nearest neighbor algorithm.The experiment in speech emotion recognition is implemented and the results show the effectiveness of the improved algorithm.
出处
《微电子学与计算机》
CSCD
北大核心
2015年第3期59-62,共4页
Microelectronics & Computer
基金
国家自然科学基金项目(61273266)
关键词
语音情感识别
模糊类别隶属度
模糊K近邻
speech emotion recognition
fuzzy class membership function
FKNN(fuzzy K-Nearest Neighbor)
