Table 5 Maximum classification accuracy values for the different combinations of classifiers and types of features.

From: A data mining approach using cortical thickness for diagnosis and characterization of essential tremor

 

Volume

Thickness

Roughness

All

Accuracy

Number of features used

Accuracy

Number of features used

Accuracy

Number of features used

Accuracy

Number of features used

Naive Bayes

0.6944

19

0.6944

37

0.8056

10

0.8056

62

Support Vector Machine

0.7500

31

0.6944

23

0.6667

36

0.5833

62

Rule

0.6667

26

0.7500

8

0.8056

36

0.6111

14

K-Nearest Neighbor

0.7222

6

0.7222

17

0.6667

33

0.7222

14

Artificial Neural Network

0.7500

6

0.6389

28

0.7222

10

0.7500

67

Average

0.7167

17.60

0.70

22.60

0.7333

25.00

0.6944

43.80

Standard deviation

0.0362

11.4149

0.0412

10.9681

0.0697

13.7477

0.0942

27.280

  1. Accuracy range is between 0 and 1. There was no significant difference among average maximum values (p < 0.05, Student’s t-test, Bonferroni-corrected).
Back to article page