Skip to main content

Table 4 Number of times different methods achieve the best results with NB classifier

From: A novel method of constrained feature selection by the measurement of pairwise constraints uncertainty

Dataset

LS

GCNC

FGUFS

FS

FAST

FJMI

PCA

PCFS

SPECTF

1

1

1

0

0

2

0

5

SpamBase

0

1

0

1

1

1

0

6

Sonar

0

1

0

0

1

0

1

7

Arrhythmia

0

1

1

0

0

1

0

7

Madelon

0

1

1

0

1

1

0

6

Isolet

0

1

1

1

1

1

0

5

Multiple features

0

2

1

1

1

1

0

4

Colon

0

1

0

0

0

2

1

6

Average

0.12

1.12

0.62

0.37

0.62

1.12

0.25

5.75