Table 11. Comparison of overall success rate on the four benchmark datasets.

		Algorithm	Representation and method	OSR (%)
Dataset 1		SubNucPred [33]	SSLD and AAC based on SVM by Jackknife test	81.46
		Effective Fusion Representations [4]	DipPSSM with LDA based on KNN by 10-fold cross-validation	≈97
			PseAAPSSM with LDA based on KNN by 10-fold cross-validation	≈84
		The proposed method: CoPSSM with intelligent KLDA based on DGGA	CoPSSM with KLDA based on KNN and Jackknife test	87.44
Dataset 2		Nuc-PLoc [7]	Fusion of PsePSSM and PseAAC based on Ensemble classifier by Jackknife test	67.4
		Effective Fusion Representations [4]	DipPSSM with LDA based on KNN by 10-fold cross-validation	95.94
			PseAAPSSM with LDA based on KNN by 10-fold cross-validation	88.1
		The proposed method: CoPSSM with intelligent KLDA based on DGGA	CoPSSM with KLDA based on KNN and Jackknife test	90.34
Dataset 3	Gneg-PLoc [34]		Fusion of GO approach and PseAAC based on Ensemble classifier by Jackknife test	87.3
	Nonlinear dimensionality reduction method [12]		Fusion of PSSM and PseAAC with KLDA based on KNN by Jackknife test	98.77
	The proposed method: CoPSSM with intelligent KLDA based on DGGA		CoPSSM with KLDA based on KNN and Jackknife test	92.34
Dataset 4	Gneg-PLoc [34]		Fusion of GO approach and PseAAC based on Ensemble classifier by Independent test	89.3
	The proposed method: CoPSSM with intelligent KLDA based on DGGA		CoPSSM with KLDA based on KNN and Independent test	94.71