Biomedical Engineering Reference
In-Depth Information
For example, after analysing the different
solutions provided by one exaction of the classi-
fication task with Hybrid Two-Population Genetic
Algorithm three valid (and similar) models were
obtained (see Table 6). Furthermore the results
were clearly superior to those obtained with the
previous alternatives. Besides, it was observed
that the solutions concentrate along specific spec-
tral areas (around the 88 wave number); It would
not be possible with the previous approaches.
Genetic Algorithm with Division into
Species
As in the previous section, several classification
models (all of them valid) can be analysed after an
execution of the classification algorithm, therefore
allowing the extraction of more information about
the areas where solutions are focused.
Table 7 shows the different solutions obtained
in one run of the method. Each solution is provided
by different species into the GA. In this search
there were different executions: for selecting
Table 7. Genetic algorithm with division into species
Low Concentrations
Low Concentrations
Training (134)
Training (134)
Validation (39)
Validation (39)
Comercial (2)
Comercial (2)
Run1: Selected Variables
[88 97]
Run1: Selected Variables
[88 97]
131 (957.76%)
131 (957.76%)
32 (82.05%)
32 (82.05%)
1 (50%)
1 (50%)
Run1: Selected Variables
[96 124]
Run1: Selected Variables
[96 124]
132 (98.51%)
132 (98.51%)
30(76.92%)
30(76.92%)
2 (100%)
2 (100%)
Run1: Selected Variables
[3 166]
Run1: Selected Variables
[3 166]
130 (97.01%)
130 (97.01%)
28 (71.79%)
28 (71.79%)
0 (0%)
0 (0%)
ANN Configuration for low concentrations
Topology: 2/ 10 / 60 / 7
ANN Configuration for low concentrations
Topology: 2/ 10 / 60 / 7
learning rate: 0.001 stop criterion: mse=2 or epochs=500.000
learning rate: 0.001 stop criterion: mse=2 or epochs=500.000
86
86
43
43
21
21
High Concentrations
High Concentrations
Run1: Selected Variables
[9 171]
Run1: Selected Variables
[9 171]
86 (100%)
86 (100%)
40 (93.02%)
40 (93.02%)
18 (85.71%)
18 (85.71%)
Run1: Selected Variables
[12 15]
Run1: Selected Variables
[12 15]
83 (96.51%)
83 (96.51%)
34 (79.07%)
34 (79.07%)
18 (85.71%)
18 (85.71%)
Run1: Selected Variables
[161 165]
Run1: Selected Variables
[161 165]
83 (96.51%)
83 (96.51%)
33 (76.74%)
33 (76.74%)
17 (80.95%)
17 (80.95%)
ANN Configuration
Topology: 2 / 10 / 60 / 5 learning rate: 0.001 stop criterion: mse=2 or epochs=500.000
ANN Configuration
Topology: 2 / 10 / 60 / 5 learning rate: 0.001 stop criterion: mse=2 or epochs=500.000
Search WWH ::
Custom Search