Research Article
Prediction of Concrete Compressive Strength by Evolutionary Artificial Neural Networks
Table 4
Results obtained from MLP, FF, RBF, and TLRN optimal models in the training, validation, and test stages for compressive strength of concrete.
| Test | Validation concurrent with training | Train |
Model |
Number | Best fitting line in testing phase | Performance criteria | Best fitting line in validation phase | Performance criteria | Best fitting line in training phase | Performance criteria | Equation | | NMSE | | Equation | | NMSE | | Equation | | NMSE | |
| | 0.899 | 0.106 | 0.948 | | 0.935 | 0.070 | 0.967 | | 0.910 | 0.090 | 0.954 | MLP | 1 | | 0.843 | 0.179 | 0.918 | | 0.916 | 0.126 | 0.957 | | 0.794 | 0.206 | 0.891 | FF | 2 | | 0.711 | 0.380 | 0.843 | | 0.783 | 0.379 | 0.885 | | 0.734 | 0.381 | 0.857 | RBF | 3 | | 0.473 | 0.705 | 0.688 | | 0.317 | 1.180 | 0.563 | | 0.411 | 0.813 | 0.641 | TLRN | 4 |
|
|