TABLE 2.

Accuracy, sensitivity, and specificity of prediction models for different data sets compared over three machine learning approaches namely, random forests (ntree = 800, mtry = 18), linear and radial support vector machines for ROSMAP data

Random forests
		HV		MCI		AD
Data set	Accuracy	Sensitivity	Specificity	Sensitivity	Specificity	Sensitivity	Specificity
1	0.9605	0.9947	0.9545	0.9444	0.9648	0.5	1
2	0.9407	0.989	0.875	0.8254	0.9789	0.77778	0.9918
3	0.9407	0.9838	0.8971	0.8413	0.9737	0.6	0.9879
4	0.9328	0.9944	0.863	0.8525	0.9635	0.41667	1
5	0.9644	1	0.918	0.8868	0.985	0.625	0.9959
6	0.9407	0.9944	0.8378	0.8261	0.9837	0.6	1
7	0.9605	0.9945	0.9143	0.8923	0.984	0.6	0.996
Linear support vector machine
1	0.8656	0.9358	0.7424	0.6667	0.9246	0.6667	0.9917
2	0.8538	0.9282	0.7361	0.6984	0.9105	0.44444	0.9959
3	0.8261	0.9027	0.6618	0.6032	0.9	0.8	0.9919
4	0.8617	0.95	0.726	0.6721	0.9219	0.5	1
5	0.8972	0.9583	0.7541	0.6981	0.95	0.75	0.9959
6	0.8656	0.9721	0.6486	0.6087	0.962	0.6	0.996
7	0.8538	0.9344	0.7286	0.6462	0.9255	0.6	0.9839
Radial support vector machine
1	0.8775	0.9572	0.7576	0.7037	0.9246	0.41667	1
2	0.8696	0.9558	0.7361	0.6667	0.9368	0.55556	0.9918
3	0.8854	0.9622	0.7059	0.6825	0.9526	0.6	1
4	0.8656	0.9722	0.7123	0.6557	0.9323	0.3333	1
5	0.8854	0.9531	0.7213	0.6792	0.94	0.625	1
6	0.8775	0.9888	0.6622	0.6377	0.9674	0.2	1
7	0.8775	0.9563	0.7286	0.6923	0.9415	0.4	0.996

Abbreviations: AD, Alzheimer's disease; HV, healthy; MCI, mild cognitive impairment; ROSMAP, the Rush Religious Orders Study and Rush Memory and Aging Project