Table 4.
Comparison of the removal efficiency of 20 neurons in the hidden layer by the artificial neural network (ANN) model developed with the Levenberg–Marquardt algorithm for Pb(II) adsorption.
| No. of Neurons | R2 Value | MSE Value |
|---|---|---|
| 1 | 0.9241 | 0.0045 |
| 2 | 0.9572 | 0.0026 |
| 3 | 0.8901 | 0.0075 |
| 4 | 0.9323 | 0.0061 |
| 5 | 0.9845 | 0.0022 |
| 6 | 0.9744 | 0.0032 |
| 7 | 0.9648 | 0.0024 |
| 8 | 0.9778 | 0.0026 |
| 9 | 0.9799 | 0.001 |
| 10 | 0.9915 | 0.0028 |
| 11 | 0.9241 | 0.0035 |
| 12 | 0.9543 | 0.0025 |
| 13 | 0.9727 | 0.0041 |
| 14 | 0.9815 | 0.0021 |
| 15 | 0.9450 | 0.0016 |
| 16 | 0.9855 | 0.0029 |
| 17 | 0.9712 | 0.0038 |
| 18 | 0.9487 | 0.0015 |
| 19 | 0.9876 | 0.0027 |
| 20 | 0.9267 | 0.0043 |