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

Table 3

Prediction accuracies (variances in the brackets) of the proposed method for four data sets and comparison with other reported results.

Data set	Method	Prediction accuracy (%)
		All-α	All-β			Overall
25PDB	AADP-PSSM [44]	69.1	83.7	85.6	35.7	70.7
	AAC-PSSM-AC [45]	85.3	81.7	73.7	55.3	74.1
	SCPRED [32, 33]	92.6	80.1	74.0	71.0	79.7
	MODAS [34]	92.3	83.7	81.2	68.3	81.4
	RKS-PPSC [47]	92.8	83.3	85.8	70.1	82.9
	Ding et al. [46]	95.0	81.3	83.2	77.6	84.3
	Xia et al. [49]	92.6	72.5	71.7	71.0	77.2
	Zhang et al. [36]	95.7	80.8	82.4	75.5	83.7
	Ding et al. [48]	91.7	80.8	79.8	64.0	79.0
	Zhang et al. [16]	94.4	83.3	83.5	73.2	83.6
	This paper	94.8	95.3	89.9	85.7	91.5
D640	SCEC [38]	73.9	61.0	81.9	33.9	62.3
	SCPRED [32, 33]	90.6	81.8	85.9	66.7	80.8
	RKS-PPSC [47]	89.1	85.1	88.1	71.4	83.1
	Ding et al. [46]	92.8	88.3	85.9	66.1	82.7
	Zhang et al. [16]	92.0	81.8	87.6	74.3	83.6
	Kong et al. [20]	94.2	80.5	87.6	77.2	84.5
	This paper	97.1	92.8	97.1	80.7	91.7
FC699	SCPRED [32, 33]	—	—	—	—	87.5
	11 features [35]	97.7	88.0	89.1	84.2	89.6
	Kong et al. [20]	96.2	90.7	96.3	69.5	92.0
	This paper	97.7	98.5	95.2	96.3	96.7
1189	AADP-PSSM [44]	69.1	83.7	85.6	35.7	70.7
	AAC-PSSM-AC [45]	80.7	86.4	81.4	45.2	74.6
	SCPRED [32, 33]	89.1	86.7	89.6	53.8	80.6
	MODAS [34]	92.3	87.1	87.9	65.4	83.5
	RKS-PPSC [47]	89.2	86.7	82.6	65.6	81.3
	Zhang et al. [36]	92.4	84.4	84.4	73.4	83.6
	Ding et al. [46]	89.2	88.8	85.6	58.5	81.2
	Zhang et al. [16]	91.5	86.7	82.0	66.4	81.8
	Kong et al. [20]	91.9	84.4	85.3	72.2	83.5
	This paper	94.6	89.7	82.1	81.7	86.6