Deep Learning Structure for Cross-Domain Sentiment Classification Based on Improved Cross Entropy and Weight

Table 10

Comparison of classification results on different network models.

Dataset	Network model	Accuracy (%)	Loss (%)	Time (s)
English dataset	LSTM	80.98	0.8170	1187
	LSTM-BO	82.14	0.2267	1136
	CNN	88.34	0.5588	1374
	CNN-BO	88.74	0.1637	1172
Chinese dataset	LSTM	86.92	0.6820	1425
	LSTM-BO	87.51	0.1705	1263
	CNN	88.84	0.5874	1523
	CNN-BO	88.91	0.2231	1395