计算机工程与应用 ›› 2018, Vol. 54 ›› Issue (16): 113-118.DOI: 10.3778/j.issn.1002-8331.1704-0361

• 模式识别与人工智能 • 上一篇    下一篇

改进GWO优化SVM的语音情感识别研究

陈  闯1,RYAD Chellali1,邢  尹2   

  1. 1.南京工业大学 电气工程与控制科学学院,南京 211816
    2.桂林理工大学 测绘地理信息学院,广西 桂林 541004
  • 出版日期:2018-08-15 发布日期:2018-08-09

Research on speech emotion recognition based on improved GWO optimized SVM

CHEN Chuang1, RYAD Chellali1, XING Yin2   

  1. 1.College of Electrical Engineering and Control Science, Nanjing Tech University, Nanjing 211816, China
    2.College of Geomatics and Geoinformation, Guilin University of Technology, Guilin, Guangxi 541004, China
  • Online:2018-08-15 Published:2018-08-09

摘要: 语音情感识别日益受到人们的关注,在社会生活中发挥着重要作用。为了提高语音情感的识别率,提出一种改进的灰狼算法(Grey Wolf Optimizer,GWO)优化支持向量机(Support Vector Machine,SVM)的分类模型(IGWO-SVM)。介绍了灰狼算法的基本理论;嵌入选择算子和引入非线性收敛因子来提升IGWO的寻优性能;采用IGWO优化SVM参数,进而建立语音情感的分类模型。通过10个基准测试函数的仿真实验,验证了IGWO性能优于GWO。对于参比模型,IGWO-SVM模型能够有效提高语音情感的识别率。

关键词: 语音情感识别, 灰狼算法(GWO), 支持向量机(SVM), 选择算子, 收敛因子

Abstract: Speech emotion recognition has been paid more and more attention and plays an important role in social life. In order to improve recognition rate, IGWO-SVM is proposed. Firstly, the fundamental theory of GWO(Grey Wolf Optimizer) is presented. Then the selection operator and nonlinear convergence factor are introduced to improve performance of IGWO(Improved GWO). Finally, IGWO is used to optimize SVM parameters, and then the classification model is established. The simulation results of 10 benchmark functions show that IGWO outperforms GWO. Compared with the reference models, IGWO-SVM model can effectively improve the recognition rate of speech emotion.

Key words: speech emotion recognition, Grey Wolf Optimizer(GWO), Support Vector Machine(SVM), selection operator, convergence factor