Year: 2025
Type: article
Abstract: The generalized extreme value (GEV) distribution has been widely used in the prediction of extreme events in the most diverse areas of knowledge, being possible calculate accurately extreme quantiles of maximum data. Although the mean and variance of the GEV can be written as a function of the parameters, making a reparametrization where the parameters themselves become the mean and the standard deviation produces interpretive advantages, especially in models where these parameters are written as a function of regression models, being able to directly show how covariates affect the mean and standard deviation of the data. The proposed modelsare fitted by applying the Bayesian paradigm and assuming normal prior distribution for theregression parameters. This work proposes a reparameterization of the GEV distribution as a function of the mean and standard deviation, applied to minimum humidity data in the Northeast region of Brazil. Results obtained showed the advantage of the proposed methodology, where we can punctuate in detail how seasonal and locality variables affect the average and the variance of the minimum humidity.
Source: Mathematics
DOI: 10.3390/math13182993
Language: English
Subfield: Statistics and Probability
Field: Mathematics
Domain: Physical Sciences
SDG: Climate action
Open Access status: gold
Estimated APC paid: $3,137
OpenAlex ID: W4414239185