Table 11 Cross-validation results for classifiers based on the ANN model and the BERT model.

Model name	Reference model				STRATEGY 1				STRATEGY 2
	Accuracy	SD	Emotica	SD	Accuracy	SD	Emotica	SD	Accuracy	SD	Emotica	SD
	[%]	[%]	[%]	[%]	[%]	[%]	[%]	[%]	[%]	[%]	[%]	[%]
Training data
ANN-PCA	93.62	0.18	77.68	0.49	97.47	0.34	89.47	1.30	98.88	0.09	94.92	0.33
ANN-LSA	95.35	0.19	82.99	0.87	97.82	0.03	90.82	0.40	99.45	0.02	97.08	0.28
ANN-ICA	95.50	0.10	83.44	0.31	96.01	0.03	84.65	0.22	97.72	0.14	90.46	0.71
BERT	96.27	0.10	86.32	0.59	99.90	0.07	99.45	0.35	99.77	0.04	98.70	0.28
Testing data
ANN-PCA	93.62	0.37	77.68	0.99	93.97	0.52	78.58	1.74	93.97	0.32	79.03	0.70
ANN-LSA	95.35	0.39	82.98	1.74	95.38	0.48	82.72	1.34	95.38	0.54	83.79	1.77
ANN-ICA	95.50	0.20	83.44	0.62	95.68	0.28	83.71	0.89	95.52	0.21	83.44	1.10
BERT	96.54	0.16	87.41	0.24	96.56	0.35	87.72	0.41	96.49	0.44	86.92	2.06

Quick links

Search