Table 3.

Prediction performance of various model architectures evaluated based on testing dataset and all datasets

Measures	Testing dataset			All datasets
	CNN	BGRU	DeepSec	CNN	BGRU	DeepSec
Accuracy	0.806068	0.855145	0.871139	0.825292	0.811404	0.871345
Recall	0.782884	0.911252	0.872120	0.777778	0.931287	0.887427
Precision	0.875502	0.687112	0.868200	0.872807	0.691520	0.855263
F-measure	0.858212	0.904143	0.910294	0.816577	0.831593	0.873381
MCC	0.587026	0.608209	0.691481	0.653542	0.64152	0.743075
AUC	0.912392	0.897955	0.940572	0.906867	0.903960	0.941104

Note: The highest scores are in bold.