Table 1 Sample Table
Class
|
Attributes
|
A
|
B
|
C
|
D
|
X
|
4.6
|
3.4
|
1.4
|
0.3
|
X
|
5.0
|
4.9
|
1.5
|
0.2
|
X
|
5.8
|
4.0
|
1.2
|
0.2
|
X
|
5.1
|
3.7
|
1.5
|
0.4
|
X
|
4.3
|
3.0
|
1.1
|
0.1
|
Y
|
6.5
|
2.8
|
4.6
|
1.5
|
Y
|
4.9
|
2.4
|
3.3
|
1.9
|
Y
|
5.1
|
2.7
|
3.9
|
1.8
|
Y
|
5.0
|
2.0
|
3.5
|
1.0
|
Y
|
6.3
|
3.3
|
6.0
|
2.5
|
Z
|
5.6
|
2.8
|
4.9
|
2.0
|
Z
|
6.2
|
2.8
|
4.8
|
1.8
|
Z
|
5.9
|
3.0
|
5.1
|
2.5
|
Z
|
5.8
|
2.7
|
1.5
|
1.9
|
Z
|
6.9
|
3.2
|
5.7
|
2.3
|
Table 2 Different values of and for sample table 1
Class
|
First Generation of CIDE
|
Second Generation of CIDE
|
|
|
|
|
|
|
X
|
4.3
|
15.33
|
6.12
|
2.00
|
2.42
|
29.16
|
4.8
|
25.00
|
3.84
|
1.40
|
1.93
|
34.09
|
5.6
|
29.00
|
3.33
|
2.80
|
3.33
|
23.07
|
4.7
|
12.75
|
7.27
|
2.20
|
2.46
|
28.84
|
4.2
|
43.00
|
2.27
|
1.90
|
2.72
|
26.82
|
Y
|
5.0
|
4.33
|
18.75
|
1.80
|
1.64
|
37.83
|
3.0
|
2.57
|
27.94
|
0.90
|
1.37
|
42.10
|
3.3
|
2.83
|
26.08
|
1.20
|
1.44
|
40.90
|
4.0
|
5.00
|
16.66
|
1.50
|
1.75
|
36.36
|
3.8
|
2.52
|
28.40
|
2.70
|
1.81
|
35.48
|
Z
|
3.6
|
2.80
|
23.25
|
2.10
|
1.75
|
36.36
|
4.4
|
3.44
|
22.50
|
2.00
|
1.71
|
36.84
|
3.4
|
2.36
|
29.76
|
2.10
|
1.70
|
37.03
|
4.3
|
3.86
|
20.54
|
0.80
|
1.42
|
41.30
|
4.6
|
3.00
|
25.00
|
2.50
|
1.78
|
35.95
|
Table 3 Characteristics of classification datasets
Datasets
|
Number of Examples
|
Number of Attributes
|
Number of Classes
|
Iris
|
150
|
4
|
3
|
Wine
|
178
|
13
|
3
|
New-Thyroid
|
215
|
5
|
3
|
Breast
|
286
|
9
|
2
|
Wisconsin
|
683
|
9
|
2
|
Splice Junction Dataset
|
3190
|
61
|
3
|
Table 4 Parameter specifications
Algorithms
|
Parameters
|
DE
|
Population Size=40, Iteration=500, F=0.5, CR=0.9, Reduction Rate=0.95/0.99
|
SADE
|
Population Size=40, Iteration=500, Learning Period=50 and 100, Reduction Rate=0.95/0.99
|
JADE
|
Population Size=40, Iteration=500,p (percentage (per unit) of individuals that are considered in the mutation strategy) = 0.05, c=0.1, Reduction Rate=0.95/0.99
|
DEGL
|
Population Size=40, Iteration=500, F=0.8, CR=0.9, Weight Factor=0.0, Weight Scheme= Exponential, Adaptive, Random, and Linear, Reduction Rate=0.95/0.99
|
SFLSDE
|
Population Size=40, Iteration=500, iterSFGSS=8,iterSFHC=20, =0.1,=0.9, Reduction Rate=0.95/0.99
|
Table 5 Classification accuracies of different DE models
Algorithms
|
Testing accuracies of different datasets in %age
|
Iris
|
Wine
|
New-Thyroid
|
Breast
|
Wisconsin
|
Splice
|
DE/Rand/1/Bin
|
74.31
|
71.78
|
74.14
|
74.08
|
73.26
|
74.13
|
DE/Best/1/Bin
|
74.98
|
73.32
|
75.32
|
76.04
|
75.79
|
72.67
|
DE/RandToBest/1/Bin
|
73.63
|
74.58
|
75.63
|
75.38
|
73.62
|
73.24
|
DE/Best/2/Bin
|
75.76
|
72.46
|
74.41
|
73.15
|
74.27
|
72.22
|
DE/Rand/2/Bin
|
74.93
|
76.01
|
71.93
|
74.37
|
76.13
|
75.41
|
DE/RandToBest/2/Bin
|
75.37
|
74.51
|
77.38
|
80.03
|
78.69
|
74.52
|
SADE
(Learning Period=50)
|
78.04
|
80.02
|
79.68
|
80.08
|
79.41
|
77.04
|
SADE
(Learning Period=100)
|
77.13
|
74.13
|
80.21
|
83.14
|
81.07
|
76.02
|
JADE
|
75.84
|
76.77
|
83.98
|
78.48
|
77.56
|
76.32
|
DEGL (Exponential)
|
77.83
|
78.93
|
76.19
|
79.67
|
78.04
|
75.84
|
DEGL (Adaptive)
|
75.13
|
76.12
|
79.10
|
77.25
|
77.53
|
75.72
|
DEGL (Random)
|
76.05
|
74.07
|
77.04
|
78.72
|
77.41
|
74.34
|
DEGL (Linear)
|
75.82
|
76.51
|
75.89
|
78.42
|
76.83
|
75.94
|
SFLSDE/Rand/1/Bin
|
79.97
|
74.81
|
78.83
|
82.72
|
80.06
|
78.24
|
SFLSDE/RandToBest/1/Bin
|
81.43
|
79.17
|
80.41
|
81.88
|
80.32
|
79.02
|
1-R
|
85.94
|
79.96
|
83.42
|
82.64
|
78.88
|
79.22
|
CIDE
|
86.66
|
78.65
|
83.72
|
84.38
|
79.89
|
80.32
|
Table 6 Accuracy of standard deviation for different dataset
Algorithms
|
Accuracy of standard deviation in %age
|
Iris
|
Wine
|
New-Thyroid
|
Breast
|
Wisconsin
|
Splice
|
Average
|
DE/Rand/1/Bin
|
3.17
|
3.84
|
4.79
|
4.11
|
5.81
|
6.01
|
4.62
|
DE/Best/1/Bin
|
4.17
|
3.76
|
4.24
|
5.10
|
3.22
|
4.82
|
4.21
|
DE/RandToBest/1/Bin
|
2.18
|
3.97
|
4.36
|
5.21
|
4.88
|
3.98
|
4.09
|
DE/Best/2/Bin
|
3.25
|
3.10
|
4.56
|
5.03
|
4.78
|
5.11
|
4.30
|
DE/Rand/2/Bin
|
6.13
|
4.11
|
3.73
|
4.06
|
2.91
|
4.02
|
4.16
|
DE/RandToBest/2/Bin
|
4.02
|
3.92
|
4.13
|
2.98
|
3.07
|
3.26
|
3.56
|
SADE
(Learning Period=50)
|
3.81
|
2.97
|
4.03
|
3.62
|
3.74
|
3.83
|
3.66
|
SADE
(Learning Period=100)
|
3.26
|
3.81
|
4.03
|
2.98
|
3.24
|
3.64
|
3.49
|
JADE
|
4.12
|
3.81
|
5.11
|
2.89
|
3.04
|
4.02
|
3.83
|
DEGL (Exponential)
|
3.04
|
2.77
|
3.41
|
4.03
|
3.74
|
3.21
|
3.36
|
DEGL (Adaptive)
|
2.11
|
3.04
|
2.84
|
2.96
|
3.68
|
3.53
|
3.02
|
DEGL (Random)
|
3.02
|
4.01
|
2.79
|
4.11
|
3.24
|
4.21
|
3.56
|
DEGL (Linear)
|
2.89
|
2.46
|
2.04
|
3.68
|
2.86
|
3.47
|
2.90
|
SFLSDE/Rand/1/Bin
|
4.18
|
3.21
|
3.73
|
2.17
|
2.42
|
4.28
|
3.33
|
SFLSDE/RandToBest/1/Bin
|
3.47
|
2.81
|
3.03
|
1.90
|
2.14
|
3.72
|
2.84
|
1-R
|
2.98
|
3.13
|
2.32
|
3.88
|
4.88
|
5.32
|
3.75
|
CIDE
|
1.41
|
3.28
|
1.94
|
2.63
|
1.87
|
3.24
|
2.39
|
Table 7 Average ranking of the algorithms for datasets (Friedman Aligned-Ranks)
Algorithms
|
Average ranking of the algorithms for datasets
|
Iris
|
Wine
|
New-Thyroid
|
Breast
|
Wisconsin
|
Splice
|
Average
|
DE/Rand/1/Bin
|
16
|
17
|
16
|
16
|
17
|
14
|
16
|
DE/Best/1/Bin
|
14
|
15
|
14
|
13
|
14
|
16
|
14.33
|
DE/RandToBest/1/Bin
|
17
|
11
|
13
|
14
|
16
|
15
|
14.33
|
DE/Best/2/Bin
|
11
|
16
|
15
|
17
|
15
|
17
|
15.16
|
DE/Rand/2/Bin
|
15
|
9
|
17
|
15
|
13
|
11
|
13.33
|
DE/RandToBest/2/Bin
|
12
|
12
|
9
|
7
|
7
|
12
|
9.83
|
SADE
(Learning Period=50)
|
5
|
1
|
6
|
6
|
6
|
5
|
4.83
|
SADE
(Learning Period=100)
|
7
|
13
|
5
|
2
|
1
|
7
|
5.83
|
JADE
|
9
|
6
|
1
|
10
|
9
|
6
|
6.83
|
DEGL (Exponential)
|
6
|
4
|
11
|
8
|
8
|
9
|
7.66
|
DEGL (Adaptive)
|
13
|
8
|
7
|
12
|
10
|
10
|
10.00
|
DEGL (Random)
|
8
|
14
|
10
|
9
|
11
|
13
|
10.83
|
DEGL (Linear)
|
10
|
7
|
12
|
11
|
12
|
8
|
10.00
|
SFLSDE/Rand/1/Bin
|
4
|
10
|
8
|
3
|
3
|
4
|
5.33
|
SFLSDE/RandToBest/1/Bin
|
3
|
3
|
4
|
5
|
2
|
3
|
3.33
|
1-R
|
2
|
2
|
3
|
4
|
5
|
2
|
3
|
CIDE
|
1
|
5
|
2
|
1
|
4
|
1
|
2.33
|
Table 8 Details of attacks of labeled records
Category of Attack
|
Name of Attack
|
Normal
|
Normal
|
DoS
|
Neptune, Smurf, Pod, Teardrop, Land, back
|
Probe
|
Portsweep, Ipsweep, Nmap, satan
|
R2L
|
Guesspassword, Ftpwrite, Imap, Phf, Multihop, Warezmaster, Warezclient
|
U2R
|
Bufferoverflow, LoadModule, Perl, Rootkit
|
Table 9 Set of features of KDDCup’99 dataset
Feature No.
|
Feature Name
|
Feature No.
|
Feature Name
|
Feature No.
|
Feature Name
|
Feature No.
|
Feature Name
|
1
|
duration (c)
|
11
|
num_failed_logins (c)
|
21
|
is_host_login (s)
|
31
|
srv_diff_host_rate (c)
|
2
|
protocol_type (s)
|
12
|
logged_in (s)
|
22
|
is_guest_login (s)
|
32
|
dst_host_count (c)
|
3
|
service (s)
|
13
|
num_compromised (c)
|
23
|
count (c)
|
33
|
dst_host_srv_count (c)
|
4
|
flag (s)
|
14
|
root_shell (c)
|
24
|
srv_count (c)
|
34
|
dst_host_same_srv_rate (c)
|
5
|
src_bytes (c)
|
15
|
su_attempted (c)
|
25
|
serror_rate (c)
|
35
|
dst_host_diff_srv-rate (c)
|
6
|
dst_bytes (c)
|
16
|
num_root (c)
|
26
|
srv_serror_rate (c)
|
36
|
dst_host_same_srv_port_ rate (c)
|
7
|
land (s)
|
17
|
num_file_creations (c)
|
27
|
rerror_rate (c)
|
37
|
dst_host_srv_diff_host_ rate (c)
|
8
|
wrong_fragment (c)
|
18
|
num_shells (c)
|
28
|
srv_serror_rate (c)
|
38
|
dst_host_serror_rate (c)
|
9
|
urgent (c)
|
19
|
num_access_files (c)
|
29
|
same_srv_rate (c)
|
39
|
dst_host_srv_serror_rate (c)
|
10
|
hot (c)
|
20
|
num_outbound_cmd (c)
|
30
|
diff_srv_rate (c)
|
40
|
dst_host_rerror_rate (c)
|
41
|
dst_host_srv_rerror_rate (c)
|
|
Table 10 Distribution of connection types in KDDCup’99 10% training
Class
|
Number of records
|
Occurrence in %age
|
Normal
|
97,277
|
19.69
|
DoS
|
391,458
|
79.24
|
Probe
|
4,107
|
0.83
|
R2L
|
1,126
|
0.23
|
U2R
|
52
|
0.01
|
Total
|
494,020
|
100.00
|
Table 11 Classification accuracies (A), False alarm rate (B) and Precision (C) of DoS+10% Normal and Probe+10% Normal datasets
Algorithms
|
Testing accuracies of different datasets in %age
|
DoS+10% Normal
|
Probe+10% Normal
|
A
|
B
|
C
|
A
|
B
|
C
|
DE/Rand/1/Bin
|
73.06
|
0.0484
|
75.01
|
73.31
|
0.0465
|
74.86
|
DE/Best/1/Bin
|
76.02
|
0.0472
|
77.34
|
73.98
|
0.0478
|
75.04
|
DE/RandToBest/1/Bin
|
74.28
|
0.0501
|
75.82
|
72.53
|
0.0497
|
74.02
|
DE/Best/2/Bin
|
73.25
|
0.0501
|
75.12
|
74.66
|
0.0523
|
75.73
|
DE/Rand/2/Bin
|
73.37
|
0.0411
|
74.92
|
73.83
|
0.0465
|
75.12
|
DE/RandToBest/2/Bin
|
79.02
|
0.0441
|
81.04
|
74.27
|
0.0439
|
75.96
|
SADE
(Learning Period=50)
|
79.06
|
0.0396
|
80.34
|
77.04
|
0.0412
|
78.87
|
SADE
(Learning Period=100)
|
82.24
|
0.0389
|
83.57
|
76.13
|
0.0401
|
77.92
|
JADE
|
77.37
|
0.0412
|
78.92
|
74.74
|
0.0398
|
75.11
|
DEGL (Exponential)
|
78.56
|
0.0432
|
79.84
|
76.73
|
0.0399
|
77.03
|
DEGL (Adaptive)
|
76.25
|
0.0411
|
77.87
|
74.13
|
0.0403
|
75.76
|
DEGL (Random)
|
77.62
|
0.0451
|
78.97
|
75.05
|
0.0413
|
76.89
|
DEGL (Linear)
|
77.42
|
0.0441
|
79.02
|
74.72
|
0.0423
|
75.41
|
SFLSDE/Rand/1/Bin
|
81.62
|
0.0398
|
82.98
|
78.87
|
0.0401
|
79.32
|
SFLSDE/RandToBest/1/Bin
|
80.78
|
0.0398
|
82.02
|
80.33
|
0.0376
|
81.89
|
1-R
|
81.53
|
0.0399
|
82.38
|
84.84
|
0.0422
|
85.56
|
CIDE
|
83.68
|
0.0387
|
84.56
|
85.57
|
0.0343
|
86.87
|
Table 12 Classification accuracies (A), False alarm rate (B) and Precision (C) of R2L+10% Normal and U2R+10% Normal datasets
Algorithms
|
Testing accuracies of different datasets in %age
|
R2L+10% Normal
|
U2R+10% Normal
|
A
|
B
|
C
|
A
|
B
|
C
|
DE/Rand/1/Bin
|
70.68
|
0.0501
|
72.15
|
73.14
|
0.0502
|
75.02
|
DE/Best/1/Bin
|
72.32
|
0.0572
|
71.97
|
74.32
|
0.0532
|
75.81
|
DE/RandToBest/1/Bin
|
73.48
|
0.0504
|
74.86
|
74.53
|
0.0503
|
76.21
|
DE/Best/2/Bin
|
71.36
|
0.0506
|
72.32
|
73.31
|
0.0497
|
75.03
|
DE/Rand/2/Bin
|
75.01
|
0.0473
|
76.87
|
71.43
|
0.0501
|
73.02
|
DE/RandToBest/2/Bin
|
73.51
|
0.0465
|
75.21
|
75.36
|
0.0464
|
74.96
|
SADE
(Learning Period=50)
|
79.02
|
0.0401
|
80.12
|
78.58
|
0.0476
|
79.73
|
SADE
(Learning Period=100)
|
73.03
|
0.0489
|
75.03
|
79.21
|
0.0493
|
80.65
|
JADE
|
75.67
|
0.0476
|
76.23
|
82.88
|
0.0413
|
84.03
|
DEGL (Exponential)
|
77.83
|
0.0487
|
79.01
|
75.19
|
0.0401
|
76.89
|
DEGL (Adaptive)
|
75.12
|
0.0443
|
76.75
|
78.10
|
0.0406
|
79.93
|
DEGL (Random)
|
73.07
|
0.0501
|
76.02
|
76.04
|
0.0486
|
77.78
|
DEGL (Linear)
|
75.41
|
0.0488
|
76.94
|
74.79
|
0.0459
|
76.34
|
SFLSDE/Rand/1/Bin
|
73.71
|
0.0403
|
75.23
|
77.73
|
0.0496
|
79.21
|
SFLSDE/RandToBest/1/Bin
|
78.17
|
0.0411
|
80.12
|
79.41
|
0.0523
|
80.84
|
1-R
|
78.86
|
0.0402
|
80.23
|
82.32
|
0.0497
|
83.79
|
CIDE
|
77.65
|
0.0423
|
79.86
|
82.62
|
0.0488
|
84.03
|