Skip to main content

Table 6 Mean and standard deviation (in parentheses) of the number of features, calculated over 50 hold out runs, where the base learners are ANNs with M = 2 units in the hidden layer.

From: Improving prediction accuracy of tumor classification by reusing genes discarded during gene selection

  

Breast

Colon

Leukemia

Ovarian

Average

GA-FS

input

15564.4(2.1)

897.3(4.3)

3245.3(2.5)

10037.4(3.4)

7436.1(3.1)

discarded

8916.6(1.5)

1103.6(3.2)

3883.6(2.7)

5116.6(3.8)

4755.1(2.8)

 

H-MTL

input

12547.4(6.5)

1014.0(5.4)

4007.8(2.5)

8924.7(2.5)

6623.5(4.2)

output

4182.5(6.9)

338.0(3.2)

1335.9(3.4)

2974.9(3.4)

2207.8(4.2)

 

discarded

7751.1(5.3)

648.2(4.3)

1785.3(2.6)

3254.6(2.5)

3359.8(3.7)

 

GA-MTL

input

15624.5(2.7)

993.3(3.3)

3324.7(2.0)

10154.2(4.4)

7524.2(3.1)

output

8856.5(2.9)

1007.6(3.5)

3804.3(3.7)

4999.8(2.8)

4667.1(3.2)

 

GA-MTL-IR

input

12656.3(3.6)

877.4(4.5)

4231.6(2.9)

7895.4(3.5)

6415.2(3.6)

output

4073.6(4.3)

474.4(4.6)

1112.1(2.7)

4004.0(3.1)

2416.0(3.7)

 

discarded

7751.1(5.3)

648.2(4.3)

1785.3(2.6)

3254.6(2.5)

3359.8(3.7)

 

e-GA-MTL

input

12743.3(4.1)

884.7(5.2)

4296.4(2.9)

10235.2(3.6)

7954.4(4.0)

output

4097.4(4.3)

486.4(4.4)

1175.6(2.1)

2354.4(4.5)

2028.5(3.8)

 

discarded

7765.2(5.4)

660.0(5.1)

1796.2(2.1)

2449.2(3.6)

3167.7(4.0)

Â