Table 1.

The datasets used for training and testing of our biomedical document classification

Dataset	Number of examples
	Positive	Negative	Total
GXD	12 966	12 354	25 320
GXD-caption	1630	1696	3326