Prediction of Gamma Radiation Shielding Properties Using Machine Learning Methods

Aydın E. E., İçkecan D., Gülbiçim H., Türkan M. N.

NanoTR-19, Ankara, Türkiye, 27 - 29 Ağustos 2025, ss.1-300, (Özet Bildiri)

  • Yayın Türü: Bildiri / Özet Bildiri
  • Basıldığı Şehir: Ankara
  • Basıldığı Ülke: Türkiye
  • Sayfa Sayıları: ss.1-300
  • İstanbul Ticaret Üniversitesi Adresli: Evet

Özet

The widespread use of radiation in many areas, such as industry, medicine, agriculture and nuclear energy, poses significant risks to human health. Therefore, the need for effective shielding solutions is increasing day by day. In this context, the linear attenuation coefficient (LAC), which determines the effectiveness of materials against gamma rays, stands out as a critical parameter in environments operating with radiation. LAC is usually obtained through experimental studies or detailed Monte Carlo simulations; however, these methods can be time and cost-limiting. In this study, a dataset consisting of LAC values ​​obtained from the NIST XCOM database and covering the energy range of 0.01–10 MeV was created, and this data was used to train machine learning models. The applied models include Support Vector Regression (SVR), Kernel Ridge Regression (KRR) and k-nearest Neighbor (k-NN) algorithms; the performance of the models was evaluated with MAE, MSE and R² metrics. The trained models were tested on Praseodymium Hexaboride (PrB₆) material. The estimation results show that machine learning methods provide an effective alternative for predicting LAC values [2] ​​for new materials or materials with limited literature data. This way, the experimental load can be reduced, and significant time and cost savings can be achieved. In conclusion, this study demonstrates that machine learning is a powerful tool for accurately and rapidly estimating gamma radiation attenuation coefficients. This approach can provide scientific and industrial contributions to radiation-driven systems' design and optimization processes.