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

Table 13

Comparison of classification results on different network models.

Dataset	Model	Accuracy (%)	Loss (%)	Time (s)
English dataset	W-RNN	82.54	0.4500	1346
	W-RNN-BO	83.32	0.1430	1267
Chinese dataset	W-RNN	90.66	0.4616	1753
	W-RNN-BO	91.25	0.1405	1648