Skip to main content

Table 5 User identification results for activities G1–G7 based on chosen performance metrics

From: Identifying smartphone users based on how they interact with their phones

Gestures

Classifier

Accuracy %

F-measure

Kappa

RMSE

G1

SVM

69.75

0.699

0.6851

0.1867

Random forests

73.45

0.720

0.7238

0.1445

Bayes net

65.43

0.628

0.6403

0.1519

G2

SVM

70.46

0.745

0.6925

0.1866

Random forests

67.78

0.677

0.6646

0.1471

Bayes net

59.73

0.581

0.5808

0.1585

G3

SVM

85.90

0.858

0.8534

0.1861

Random forests

83.22

0.824

0.8254

0.1327

Bayes net

73.82

0.736

0.7277

0.1317

G4

SVM

89.87

0.900

0.8947

0.1861

Random forests

87.34

0.872

0.8684

0.1271

Bayes net

83.54

0.833

0.8289

0.1085

G5

SVM

84.17

0.838

0.8353

0.1862

Random forests

83.54

0.830

0.8288

0.1397

Bayes net

70.88

0.705

0.6972

0.1385

G6

SVM

92.15

0.918

0.9184

0.186

Random forests

92.81

0.922

0.9252

0.1218

Bayes net

90.19

0.901

0.898

0.0836

G7

SVM

92.15

0.921

0.9184

0.1861

Random forests

92.81

0.925

0.9252

0.1178

Bayes net

90.84

0.904

0.9048

0.0786