A soft computing methodology, PreFuRGe (Predictive Fuzzy Rules Generator), was applied to study atmospheric radon in relation to other meteorological variables. This algorithm is based on fuzzy clustering and generates a rule-based system that provides a qualitative model that describes the behavior of radon. Two one-year radon datasets were used to verify the consistency of the fuzzy methodology, by comparing the obtained models with published results obtained with statistical techniques. These datasets were collected in the city of Huelva, to the southwest of Europe, and were already studied in peer-reviewed publications. The proposed fuzzy methodology was able to provide results consistent with previous studies, identifying the main drivers in radon behavior, atmospheric stability and wind speed, together with a complete characterization of the daily cycle and its seasonality. PreFuRGe highlighted features that would otherwise require detailed examination combining a number of classical statistical techniques, proving to be a useful tool to characterize heterogeneous datasets.
Thesis Collection
Last Deposit
