Speech Emotion Recognition Based on PSO-optimized SVM
Abstract
Speech emotion recognition, as a significant part of man-machine interaction, has been widely used in manufacture, information industry, criminal investigation and security protection. However, recent research shows that it is hard to get a high recognition rate. This paper intends to enhance precision of recognition based on Support Vector Machine and Particle Swarm Optimization algorithm. Particle Swarm Optimization algorithm is applied to SVM's parameter selection and optimization. The experimental results on German EMO-DB database and Chinese CASIA database show that the accuracy rate using with PSO-optimized SVM reaches 90.09% and 73.5% respectively.
Keywords
Speech emotion recognition, Fourier parameter, Particle Swarm Optimization, Support Vector Machine, Frame length.
DOI
10.12783/dtcse/smce2017/12465
10.12783/dtcse/smce2017/12465
Refbacks
- There are currently no refbacks.