Speech Emotion Feature Selection and Classification Based on Random Forest
XING Yin
LIU Lilong
Abstract：Aiming at complex and high-dimensional speech emotion feature,selecting the key features shows great signifi?cance to reduce the complexity and improve the performance of the model. In this work,a fused feature selection method,based on the Fisher criterion and mean decrease Gini index in random forest is proposed. Firstly,the Fisher criterion and mean decrease Gini index evaluate the importance of all the features respectively. Secondly,a certain threshold is set to select preferable features,carry?ing out the intersection operation. Thirdly,the features are rearranged according to the order of the mean decrease Gini index. Final?ly,the optimal feature dimensions are determined by the recognition rate on validation set. The experimental results show that the fused feature selection method is effective and can improve the rate.
Keywords：feature selectionrandom forestFisher criterionGini indexspeech emotion recognition
Publication Date：2019-01-01
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：4( 539-542 )