TABLE 2.

Performance evaluation of our circLGB and other two learning-based algorithms on the CIRCdeep dataset.

Model	ACC	MCC	F1	References
circLGB	0.998	0.995	0.998	–
circDeep	0.942	0.883	0.940	(Chaabane et al., 2019)
PredcircRNA	0.806	0.611	0.811	(Pan and Xiong, 2015)

The best performance across different evaluation metrics is highlighted in bold for clarification. These highlights also apply to Tables 3, 4.