AD ALTA
JOURNAL OF INTERDISCIPLINARY RESEARCH
Attachment number 1.
Table 1: An Overview of Preserved neural networks
Source: Author
Profile
Train
Perf.
Select
Perf.
Test Perf.
Train
Error
Select
Error
Test Error
Training/Members
Inputs
Hidden
(1)
Hidden
(2)
1
MLP 1:4-33-61-4:1
0.94359
0.94524
0.94487
1.43022
1.54192
1.45348
BP10, CG20, CG0b
1
33
61
2
MLP 2:7-88-63-4:1
0.03850
0.03760
0.03832
1.02620
1.04046
1.01426
BP10, CG20, CG0b
2
88
63
3
MLP 15:15-54-66-4:1
0.94318
0.94505
0.94443
0.58611
0.59107
0.5835
BP10, CG20, CG0b
15
54
66
4
Linear 84:86-4:1
0.94443
0.94549
0.94487
0.16093
0.16396
0.16088
PI
84
0
0
5
Linear 90:98-4:1
0.94431
0.94549
0.94462
0.16085
0.16203
0.16074
PI
90
0
0
6
PNN 88:93-31997-4:1
0.94425
0.94605
0.94543
0.16236
0.16034
0.16066
88
31997
0
7
PNN 87:92-31997-4:1
0.94425
0.94605
0.94543
0.16237
0.16034
0.16066
87
31997
0
8
RBF 61:69-328-4:1
0.94387
0.94549
0.94480
0.15962
0.15862
0.15909
SS,KN,PI
61
328
0
9
RBF 61:69-359-4:1
0.94371
0.94512
0.94480
0.15964
0.15861
0.15993
SS,KN,PI
61
359
0
10
RBF 61:69-360-4:1
0.94387
0.94543
0.94505
0.15932
0.15859
0.15909
SS,KN,PI
61
360
0
- 239 -