Development of radon dispersion prediction model based on artificial neural network for Sin Quyen copper mine

Abstract

Radon is one of the most toxic natural radionuclides which occupy more thant 50% of natural radiation expose. The Sinquyen is a natural radioactive bearing mine in the North of Vietnam. The uranium is the dominate radionuclide in this mine, which is the source emit radon to surrounding when the mine is exploited. The machine learning have been applied in radon prediction still rare and limitation. The assessment as well as understanding the radon dispersion released from this mine are important targets for radiation hazard assessment. In this paper, we designed a simple one-hidden layer artificial neural network (ANN) that requires low computation cost to train and reference. Our comparison results suggested that the proposed model outperformed other benchmark methods such as two-hidden-layer ANN, Random Forest (RF), and Support Vector Machine (SVM). The results also revealed that distance and coordinates factors had a greater effect on the prediction of Radon prediction.