A0380
Title: Decoding weather interactions: Bayesian structure learning of causal effects in South Africa's climate
Authors: Mohammad Arashi - Ferdowsi University of Mashhad (Iran) [presenting]
Samaneh Nazari - Ferdowsi University of Mashhad (Iran)
Abdolnasser Sadeghkhani - North Carolina Agricultural and Technical State University (United States)
Abstract: To understand how different weather factors interact in South Africa, a method called Bayesian structure learning is used. This involves creating directed acyclic graphs (DAGs) to help us see the cause-and-effect relationships between various weather conditions. The nodes in these graphs represent important weather factors like temperature, wind speed, rain, cloud cover, and snow depth. DAGs are useful for showing how these factors depend on each other in complex situations. To learn about these relationships accurately, specific statistical models are used. Normal-gamma and normal-inverse gamma priors are applied to the analysis, which helps calculate the probabilities of different outcomes. The methods are tested using past weather data from South Africa.
