Skip to main content

Table 4 RF modeling, filtered data, approach II.

From: Data mining of high density genomic variant data for prediction of Alzheimer's disease risk

Data

RF

10-Fold

Cross

Validation

% Error

(APOEε4+ and ε4- samples)

Number of Trees

F = 6

F = 7

F = 11

F = 18

APOE & 10 GAB2 SNPs

     
 

100

 

33.3

33.2

 
 

150

 

33.4

33.5

 
 

200

 

33.4

33.2

 
 

250

 

33.2

33.2

 
 

300

 

33.3

33.2

 
 

600

 

33.4

33.5

 

19 SNPs

     
 

100

26.9

 

27.5

27.9

 

150

26.7

 

27.0

27.5

 

200

26.8

 

27.0

27.4

 

250

27.1

 

27.3

27.6

 

300

27.2

 

27.1

27.6

 

600

27.1

 

27.4

27.4

APOE & 19 SNPs

     
 

100

20.3

 

20.8

20.7

 

150

19.9

 

20.6

20.8

 

200

20.1

 

20.7

20.9

 

250

20.3

 

20.6

20.6

 

300

20.1

 

20.6

20.6

 

600

19.8

 

20.4

20.7

APOE , 10 GAB2 , & 19 SNPs

     
 

100

17.6

 

17.6

18.2

 

150

17.2

 

17.0

17.6

 

200

17.2

 

17.2

18.1

 

250

17.3

 

17.1

17.9

 

300

17.2

 

17.3

17.9

 

600

17.3

 

16.9

17.6

  1. F = number of features