Default Risk Prediction of Enterprises Based on Convolutional Neural Network in the Age of Big Data: Analysis from the Viewpoint of Different Balance Ratios
Table 8
The rank of four models on different datasets.
ID
Dataset
Rank
Dataset
Rank
Dataset
Rank
1
J
31
A
32
Ch
27
2
J 0
32
A 0
29
Ch 0
31
3
J 1
5
A 1
15
Ch 1
6
4
J 2
10
A 2
18
Ch 2
9
5
J 3
11
A 3
20
Ch 3
6
6
J 4
20
A 4
16
Ch 4
19
7
J 5
18
A 5
26
Ch 5
15
8
J 6
28
A 6
28
Ch 6
11
9
J 7
22
A 7
24
Ch 7
22
10
J 8
26
A 8
31
Ch 8
15
11
J 9
26
A 9
27
Ch 9
13
12
J 10
29
A 10
29
Ch 10
17
13
J 11
16
A 11
1
Ch 11
4
14
J 12
4
A 12
7
Ch 12
10
15
J 13
1
A 13
4
Ch 13
1
16
J 14
12
A 14
11
Ch 14
6
17
J 15
7
A 15
2
Ch 15
3
18
J 16
15
A 16
9
Ch 16
2
19
J 17
23
A 17
22
Ch 17
13
20
J 18
12
A 18
22
Ch 18
20
21
J 19
18
A 19
13
Ch 19
21
22
J 20
6
A 20
16
Ch 20
24
23
J 21
9
A 21
13
Ch 21
11
24
J 22
17
A 22
10
Ch 22
24
25
J 23
3
A 23
7
Ch 23
4
26
J 24
2
A 24
5
Ch 24
18
27
J 25
12
A 25
3
Ch 25
26
28
J 26
23
A 26
12
Ch 26
23
29
J 27
21
A 27
21
Ch 27
27
30
J 28
8
A 28
19
Ch 28
29
31
J 29
25
A 29
6
Ch 29
30
32
J 30
30
A 30
25
Ch 30
31
The bold value means that the method ranks first in some evaluation criterion. For example, CNN ranks first in terms of G-mean, accuracy, and AUC, which indicates that its comprehensive ranking also ranks first.
