Skip to main content

Table 8 Accuracy rates of k-means algorithm

From: An investigation of the predictability of the Brazilian three-modal hand-based behavioural biometric: a feature selection and feature-fusion approach

k-means

Algorithm

Keyboard

Keyboard

Touch + signature

algorithm

 

+ touch

+ signature

 

Euclidean distance

KNN

57.01

62.71

57.01

 

SVM

57.45

59.64

47.36

 

MLP

58.77

43.85

60.52

 

LIBSVM

53.07

63.59

65.35

 

WiSARD

55.70

67.54

61.40

Cityblock distance

KNN

63.15

64.91

68.85

 

SVM

59.64

64.03

50.00

 

MLP

59.21

41.22

67.98

 

LIBSVM

58.77

72.80

66.66

 

WiSARD

62.28

66.66

65.35

Correlation distance

KNN

56.14

64.03

59.21

 

SVM

49.12

59.64

46.05

 

MLP

57.01

53.07

64.03

 

LIBSVM

54.82

65.78

65.78

 

WiSARD

42.98

65.78

61.48

Cosine distance

KNN

57.45

66.22

57.89

 

SVM

57.01

60.08

46.05

 

MLP

55.70

53.94

62.71

 

LIBSVM

53.07

65.78

68.42

 

WiSARD

54.38

64.03

59.64