Skip to main content

Table 1 Quantitative results on the simulation data

From: ECFS-DEA: an ensemble classifier-based feature selection for differential expression analysis on expression profiles

Base classifier

Variable number

Confusion matrix

Positive class

Precision

Recall

F1-measure

LDA

[0,1]T

classified as

a

b

a

0.992

0.984

0.988

  

label a

123

2

b

0.984

0.992

0.988

  

label b

1

124

weighted average

0.988

0.988

0.988

kNN

[0,1]T

classified as

a

b

a

0.906

0.928

0.917

  

label a

116

9

b

0.926

0.904

0.915

  

label b

12

113

weighted average

0.916

0.916

0.916

RF

34

classified as

a

b

a

0.528

0.448

0.485

  

label a

56

69

b

0.521

0.600

0.558

  

label b

50

75

weighted average

0.524

0.524

0.522

 

[0,1]T

classified as

a

b

a

0.897

0.904

0.900

  

label a

113

12

b

0.903

0.896

0.899

  

label b

13

112

weighted average

0.900

0.900

0.899

SVM

0

classified as

a

b

a

0.467

0.400

0.431

  

label a

50

75

b

0.476

0.544

0.508

  

label b

57

68

weighted average

0.472

0.472

0.470

 

[0,1]T

classified as

a

b

a

0.909

0.960

0.934

  

label a

120

5

b

0.958

0.904

0.930

  

label b

12

113

weighted average

0.933

0.932

0.932