为提升情感识别的准确性和稳定性,研究了基于深度学习的语音情感识别优化方法.首先,探讨基于长短时记忆网络(Long Short Term Memory,LSTM)的语音情感识别的基本原理.其次,引入L2 正则化的优化方法.最后,进行实验分析,评估所提方法的性能.实验结果表明,所提方法在识别准确性和稳定性方面均取得良好的效果.
Optimization Method for Speech Emotion Recognition Based on Deep Learning
To improve the accuracy and stability of emotion recognition,a deep learning based optimization method for speech emotion recognition is studied.Firstly,explore the basic principles of speech emotion recognition based on Long Short Term Memory(LSTM)networks.Secondly,introduce the optimization method of L2 regularization.Finally,conduct experimental analysis to evaluate the performance of the proposed method.The experimental results show that the proposed method has achieved good results in recognition accuracy and stability.
NETL
NSTL
万方数据
