Faculty of Mathematics, Physics
and Informatics
Comenius University Bratislava

Nuclear Seminar - Dr. Mohammad Alem Sultani (9.10.2024)

Wednesday 9.10.2024 at 14:00, Lecture room F1/364


03. 10. 2024 10.56 hod.
By: Jaroslav Staníček

Dr. Mohammad Alem Sultani:
Machine Learning Approaches for Analyzing Factors Influencing Atmospheric Radon Variability in Bratislava, Slovakia


Abstract:
Radon (²²²Rn) is a radioactive noble gas that is produced in the decay chain of uranium-238 (²³⁸U). The study of outdoor radon is of interest due to its significant contribution to background radiation and its potential use as a tracer in atmospheric research. A variety of factors, including the radon exhalation rate, atmospheric mixing height, and meteorological conditions, contribute to the variability of radon activity concentration (RAC). This study examines the temporal fluctuations in RAC and the factors contributing to its variability in Bratislava, Slovakia, utilizing a four-year dataset of continuous measurements. The hourly, diurnal, and seasonal fluctuations of RAC were subjected to analysis and discussion. Furthermore, machine learning techniques, including artificial neural networks, automatic linear modeling, and random forest, were utilized in conjunction with conventional multivariate linear regression to investigate the factors influencing RAC variability. The analysis revealed that the primary factors affecting RAC are mixing height, radon flux, and temperature. Furthermore, the random forest exhibited superior performance compared to other approaches. This research contributes to a more profound understanding of the processes that shape the atmospheric dynamics of radon.