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

Table 4

Model parameter configuration.

Name	Description
Batch_size	Number of batch training samples
Hidden_dam	Hidden layer nodes
Embedding_dam	Word vector dimension
Filters	Filter size
Kernel_size	Number of convolution kernels
Max_features	The maximum characteristic number of word bag model
Epoch	Times of model iterations
Min_count	Minimum word frequency