[Retracted] Application of Deep Learning for Discourse Expression Recognition in Ideological and Political Education

Table 4

Results for different classes for different models.

Number of classes	Model	Accuracy %	Precision %	Recall %	F score
8	B + L + R + S + A	94.54	92.43	94.65	93.31
	B + L	93.23	91.12	93.34	92
	B + R + B	95.12	93.01	95.23	93.89
	B + A	92.13	90.02	92.24	90.9
	Proposed	96.32	94.21	96.43	95.09
15	B + L + R + S + A	93.65	92.88	93.54	92.2
	B + L	92.34	91.57	92.23	90.89
	B + R + B	94.23	93.46	94.12	92.78
	B + A	91.24	90.47	91.13	89.79
	Proposed	95.43	94.66	95.32	93.98
22	B + L + R + S + A	92.76	93.33	92.43	91.09
	B + L	91.45	92.02	91.12	89.78
	B + R + B	93.34	93.91	93.01	91.67
	B + A	90.35	90.92	90.02	88.68
	Proposed	94.54	95.11	94.21	92.87
28	B + L + R + S + A	91.87	93.78	91.32	89.98
	B + L	90.56	92.47	90.01	88.67
	B + R + B	92.45	94.36	91.9	90.56
	B + A	89.46	91.37	88.91	87.57
	Proposed	93.65	95.56	93.1	91.76
32	B + L + R + S + A	90.98	94.23	90.21	88.87
	B + L	89.67	92.92	88.9	87.56
	B + R + B	91.56	94.81	90.79	89.45
	B + A	88.57	91.82	87.8	86.46
	Proposed	92.76	96.01	91.99	90.65