From: Decision tree SVM model with Fisher feature selection for speech emotion recognition
Initial threshold
1~5%
6%
7~8%
9~11%
≥ 12%
Average recognition rate (%)
86.39
86.86
84.36
83.73
80.47