Table 2.

Comparison of average performance of 10-fold cross-validation in RefSeq lncRNA and protein-coding gene predictions using all features, chromatin data only, sequence only and feature selected features

Used features	AUC	Precision	Recall
All features	0.927	0.857	0.857
All chromatin features	0.827	0.760	0.759
All sequence features	0.876	0.797	0.796
Feature selected features	0.882	0.833	0.832

The comparisons include the AUC under ROC curves, precision and recall values.