Determination of neutron fluence-to-dose conversion coefficients by means of artificial neural networks

An Artificial Neural Network has been designed to determinate the effective dose, the ambient dose equivalent and the personal dose equivalent fluence-to-dose conversion factors using seven count rates obtained with a Bonner Sphere Spectrometer. The data of 211 neutron spectra and their respective fluence-to-dose conversion coefficients were used to train and to test ANN. The ANN was trained using the trainsec algorithm, the definitive ANN was 7:8:9:10. From the 30 set of data used to test the ANN performance the largest difference was 11% that is close to the difference obtained in neutron dosimetry.