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

Figure 18

Loss of each model in Chinese and English datasets. (a) Change in loss rate of RNN in English dataset. (b) Change in loss rate of W-RNN in English dataset. (c) Change in loss rate of RNN in Chinese dataset. (d) Change in loss rate of W-RNN in Chinese dataset.

(a)

(b)

(c)

(d)