Table 1 Results of the feature selection process, showing the sequential removal of features based on ANOVA F-test scores. The table includes the removed features, their corresponding F-test score, the number of remaining features, and the resulting model accuracy.

From: Machine learning-based classification of histological subtypes of invasive breast cancer using MRI contralateral breast texture features

Removed feature

ANOVA F-test score

Number of features

Accuracy

p-value

Correlation

0.1233

12

0.9038

7.254574e-01

Mean

0.5335

11

0.8997

4.651769e-01

Entropy

4.2328

10

0.8977

3.975375e-02

Standard deviation

5.9237

9

0.8936

1.500902e-02

Energy

9.0595

8

0.8691

2.639950e-03

Local entropy

11.0821

7

0.8548

8.845887e-04

Kurtosis

32.9575

6

0.8507

1.058667e-08

Skewness

34.9334

5

0.8466

3.888354e-09

Homogeneity

53.9667

4

0.8057

2.765401e-13

Local standard deviation

69.9366

3

0.7852

1.015149e-16

Local range

74.2891

2

0.7607

1.191515e-17

Contrast

87.4496

1

0.6605

1.890132e-20

Back to article page