Cross-Corpus Speech Emotion Recognition Based on Transfer Learning and Multi-Loss Dynamic Adjustment

Table 4

Experimental results of the use of other algorithms.

Algorithm	EB	BE	EC	CE	BC	CB	Average
PCA + SVM	50.85	33.48	28.40	27.61	33.13	43.38	36.14
DoSL [14]	50.55	33.03	35.20	33.81	39.23	53.20	40.84
TSDSL [17]	47.41	35.44	32.50	33.25	37.40	56.74	40.46
JDAR [18]	48.74	38.14	30.30	28.43	38.60	49.58	38.97
DBN + BP [28]	29.86	32.21	24.20	31.02	35.80	49.59	33.78
TLMLDA	58.93	43.16	35.40	32.74	41.00	57.11	44.72

The bold values are the highest recognition rate in each task to reflect the rationality of the TLMLDA model, because TLMLDA has obtained the best performance compared with other ablation experimental models.