SpringerPlus

Table 4 Summary of the three cytokines used in the integration approach

From: An approach to predict the risk of glaucoma development by integrating different attribute data

Base classifier			Single analysis			Analysis with sampling^*
			Accuracy	Sensitivity	Specificity	Accuracy	Sensitivity	Specificity
LDA		Genotype	0.688	0.712	0.654	0.671 ± 0.011	0.693 ± 0.015	0.639 ± 0.014
		Cytokine	0.592	0.466	0.769	0.584 ± 0.010	0.457 ± 0.012	0.763 ± 0.010
		Integrated	0.632	0.616	0.654	0.655 ± 0.022	0.611 ± 0.034	0.717 ± 0.015
SVM	linear	Genotype	0.664	0.699	0.615	0.683 ± 0.013	0.754 ± 0.023	0.584 ± 0.016
		Cytokine	0.568	0.452	0.731	0.577 ± 0.008	0.458 ± 0.012	0.745 ± 0.013
		Integrated	0.659	0.648	0.673	0.668 ± 0.014	0.640 ± 0.024	0.706 ± 0.012
	polynomial	Genotype	0.648	0.589	0.731	0.633 ± 0.010	0.539 ± 0.026	0.764 ± 0.018
		Cytokine	0.512	0.658	0.308	0.457 ± 0.012	0.275 ± 0.077	0.713 ± 0.086
		Integrated	0.656	0.521	0.846	0.624 ± 0.010	0.480 ± 0.065	0.827 ± 0.078
	RBF	Genotype	0.688	0.712	0.654	0.676 ± 0.010	0.685 ± 0.016	0.664 ± 0.013
		Cytokine	0.648	0.712	0.558	0.662 ± 0.006	0.701 ± 0.011	0.607 ± 0.020
		Integrated	0.744	0.767	0.712	0.740 ± 0.013	0.805 ± 0.020	0.650 ± 0.014
NBC		Genotype	0.640	0.671	0.596	0.630 ± 0.006	0.651 ± 0.013	0.601 ± 0.014
		Cytokine	0.624	0.479	0.827	0.621 ± 0.006	0.489 ± 0.013	0.807 ± 0.019
		Integrated	0.744	0.767	0.712	0.698 ± 0.013	0.644 ± 0.027	0.775 ± 0.051
DT		Genotype	0.536	0.342	0.808	0.562 ± 0.025	0.411 ± 0.070	0.774 ± 0.043
		Cytokine	0.624	0.904	0.231	0.605 ± 0.018	0.874 ± 0.099	0.226 ± 0.126
		Integrated	0.600	0.959	0.096	0.617 ± 0.013	0.668 ± 0.032	0.545 ± 0.040

^*These values are represented as the mean and SD of each statistics. The mean of each statistics included extremely good or bad result, especially small sampling size and few sampling repeat time.

Back to article page