Conference Paper
Using Feed Forward Neural Network to Solve Eigenvalue Problems
Table 4
Weight and bias of the network for different training algorithms.
(a) |
| Weights and bias for trainlm | Net.IW | Net.LW | Net.B{1} |
| 0.0092 | 0.2088 | 0.1319 | 0.4768 | 0.5205 | 0.9547 | 0.2503 | 0.2255 | 0.1239 | 0.3079 | 0.5672 | 0.1862 | 0.9669 | 0.9982 | 0.6465 |
|
|
(b) |
| Weights and bias for trainbfg | Net.IW | Net.LW | Net.B{1} |
| 0.2691 | 0.9831 | 0.6981 | 0.4228 | 0.3015 | 0.6665 | 0.5479 | 0.7011 | 0.1781 | 0.9427 | 0.6663 | 0.1280 | 0.4177 | 0.5391 | 0.9991 |
|
|
(c) |
| Weights and bias for trainbr | Net.IW | Net.LW | Net.B{1} |
| 0.5211 | 0.3955 | 0.9133 | 0.2316 | 0.3674 | 0.7962 | 0.4889 | 0.9880 | 0.0987 | 0.6241 | 0.0377 | 0.2619 | 0.6791 | 0.8852 | 0.3354 |
|
|