Table 6 Classwise comparative-analysis between proposed deep NNW and existing approaches34 for CCCS-CIC-AndMal-2020 dataset.

From: Multimodal malware classification using proposed ensemble deep neural network framework

Malware category

Precision

Recall

F1-Score

Accuracy

Proposed

Existing

Proposed

Existing

Proposed

Existing

Proposed

Existing

Adware

0.892

0.935

0.889

0.929

0.891

0.932

85.10

92.82

Backdoor

0.912

0.721

0.948

0.643

0.930

0.680

94.80

59.93

Banker

0.818

0.759

0.917

0.759

0.865

0.759

91.80

92.40

Dropper

0.850

0.850

0.826

0.686

0.838

0.759

82.60

63.96

FileInfector

0.778

0.909

0.880

0.789

0.826

0.845

88.00

70.31

PUA

0.968

0.677

0.972

0.682

0.970

0.679

97.20

69.29

Ransomware

0.920

0.798

0.931

0.944

0.926

0.864

93.10

91.98

Riskware

0.949

0.963

0.939

0.967

0.944

0.965

92.80

96.55

SMS

0.953

0.917

0.973

0.886

0.963

0.901

97.30

93.99

Scareware

0.326

0.836

0.909

0.764

0.480

0.799

91.70

74.32

Spyware

0.000

0.924

0.000

0.835

0.000

0.877

92.00

91.94

Trojan

0.962

0.895

0.894

0.896

0.927

0.896

87.40

89.09

Overall Results

0.799

0.841

0.865

0.813

0.850

0.825

91.40

82.99

  1. Significant values are in bold.