@article{391, author = "Olawale Basheer Akanbi", abstract = "Radon gas is a leading cause of lung cancer after smoking. According to Environmental Protection Agency (EPA) and Surgeon General, about 20,000 lung cancer deaths are recorded each year and are caused by radon gas Radon is readily available in all types of soil and moves freely into homes via cracks on the floor, hence can contaminate our drinking water. Previous studies have explored the danger of soil radon gas using various classical methods such as ordinary kriging to capture spatial attributes of the data sets to make useful predictions. In this study, two Bayesian approaches are considered namely, A Bayesian linear regression (Independent) Model which ignores the spatial dependency as the baseline model, and Bayesian Geostatistical Model with an additional spatial dependency parameter to quantify the relationships between Radon gas concentration and the covariate structures are considered. The effects of the environmental variables were investigated and the radon potential map of the study sites was produced using four picocuries (200 Bqm-3) as the threshold. The radon potential map indicated that 22 locations already have radon concentrations above the four picocuries threshold. The two approaches produced similar posterior means and standard deviations. However, the WAIC and DIC values indicated that the Bayesian Geostatistical Model provided a better fit for the data, hence justifying the inclusion of the spatial parameter.", issn = "23942894", journal = "IJASM", keywords = "Bayesian Geostatistical Model;Spatial Parameter;Deviance Information Criteria;Radon Gas Concentration;Geogenic Radon Potential", month = "May", number = "3", pages = "36-46", title = "{S}patial {A}nalysis of {S}oil {R}adon {G}as {C}oncentration in {S}outhwestern {N}igeria: {A} {B}ayesian {A}pproach", volume = "9", year = "2022", }