Normal scale mixture copula marginal regression with Box-Cox symmetric distributions (2022)
- Authors:
- USP affiliated authors: FERRARI, SILVIA LOPES DE PAULA - IME ; MEDEIROS, RODRIGO MATHEUS ROCHA DE - IME
- Unidade: IME
- Subjects: ANÁLISE DE SÉRIES TEMPORAIS; DISTRIBUIÇÕES (PROBABILIDADE)
- Keywords: Clustered Data; Dependence Structures; Log-Symmetric Distributions; Working Correlation Matrix
- Language: Inglês
- Abstract: The class of the Box-Cox symmetric distributions was recently introduced in the statistical literature. The class provides a flexible modeling framework for univariate independent positive continuous data with different levels of skewness and tail-heaviness. Additionally, the relatively easy parameter interpretation makes it attractive for regression purposes. However, more general applications may involve correlated data, such as when observations have a temporal or spatial dependence. Based on Sklar’s Theorem, the copula theory provides an approach to modeling dependence through a function (named copula) which describes how the elements of a random vector are associated. Particularly, copulas generated by scale mixtures of normal distributions allow the bivariate associations to determine the dependence structure of the random vector entirely. Moreover, they also achieve positive and negative associations without restrictions on the data dimension. This work introduces a broad class of marginal regression models to analyze correlated positive continuous data with Box-Cox symmetric marginal distributions, where a normal scale mixture copula describes the dependence. Our approach resembles the joint modeling of univariate observations of the classical generalized estimating equations model. It is possible to select one of several association structures specified in terms of nonlinear response transformations, which provides flexibility in modeling independent observations, time series, longitudinal, clustered, or spatially correlated data.
- Imprenta:
- Source:
- Título do periódico: Livro de Resumos
- Conference titles: Simpósio Nacional de Probabilidade e Estatística - SINAPE
-
ABNT
MEDEIROS, Rodrigo Matheus Rocha de e FERRARI, Sílvia Lopes de Paula. Normal scale mixture copula marginal regression with Box-Cox symmetric distributions. 2022, Anais.. São Paulo: ABE, 2022. Disponível em: https://app.eventize.com.br/upload/004449/files/Sinape2022_FINAL.pdf. Acesso em: 01 maio 2024. -
APA
Medeiros, R. M. R. de, & Ferrari, S. L. de P. (2022). Normal scale mixture copula marginal regression with Box-Cox symmetric distributions. In Livro de Resumos. São Paulo: ABE. Recuperado de https://app.eventize.com.br/upload/004449/files/Sinape2022_FINAL.pdf -
NLM
Medeiros RMR de, Ferrari SL de P. Normal scale mixture copula marginal regression with Box-Cox symmetric distributions [Internet]. Livro de Resumos. 2022 ;[citado 2024 maio 01 ] Available from: https://app.eventize.com.br/upload/004449/files/Sinape2022_FINAL.pdf -
Vancouver
Medeiros RMR de, Ferrari SL de P. Normal scale mixture copula marginal regression with Box-Cox symmetric distributions [Internet]. Livro de Resumos. 2022 ;[citado 2024 maio 01 ] Available from: https://app.eventize.com.br/upload/004449/files/Sinape2022_FINAL.pdf - Redução de vício de estimadores de máxima verossimilhança na família exponencial uniparamétrica
- Adjusted profile likelihood for two-parameter exponential family models
- Modified score test statistic having chi-squared distribution to order 'N POT.-1'
- Bartlett and Bartlett-type corrections for testing linear restrictions
- Uma análise de associação entre poluição atmosférica e saúde, usando um modelo de regressão binomial negativo
- An improved likelihood ratio test for varying dispersion in exponential family nonlinear models
- Inflated beta distributions
- Three Bartlett-type corrections for score statistics in symmetric nonlinear regression models
- Small-sample corrections for score tests in Birnbaum-Saunders regressions
- Size and power properties of some tests in the Birnbaum-Saunders regression model
Download do texto completo
Tipo | Nome | Link | |
---|---|---|---|
3186409.pdf | Direct link |
How to cite
A citação é gerada automaticamente e pode não estar totalmente de acordo com as normas