Reordering Features with Weights Fusion in Multiclass and Multiple-Kernel Speech Emotion Recognition

Table 4

Comparison on SER accuracies (%) of 5 emotions in Berlin database.

Methods	Anger	Fear	Happy	Neutral	Sad
Enhanced-SRC	98.55	83.16	57.73	70.08	96.71
Feature fusion based on MKL	81	83		65	95
RFWF MK	90.476	79.412	97.143	87.179	100