Using Recursive Feature Selection with Random Forest to Improve Protein Structural Class Prediction for Low-Similarity Sequences

Table 2

Sensitivity (Sens), specificity (Spec), and F1 of the proposed method on four data sets.

Data set	Class	Sens (%)	Spec (%)	F1 (%)
25PDB	All-α	94.81	98.29	95.02
	All-β	95.26	98.13	95.05
		89.88	95.25	86.39
		85.71	97.16	88.52
D640	All-α	97.10	97.81	94.70
	All-β	92.86	99.18	95.02
		97.18	92.87	90.05
		80.70	98.93	87.90
FC699	All-α	97.69	99.45	97.32
	All-β	98.51	99.49	98.70
		95.23	99.38	97.16
		96.34	97.68	88.27
1189	All-α	94.62	96.55	90.95
	All-β	89.80	98.50	92.63
		82.04	94.20	84.05
		81.74	92.95	79.12