KGsurv: an R package to fit the Kumaraswamy-G family of distributions for survival data

• Autor(a) principal: Caio Gabriel Barreto Balieiro
• Data de Publicação: 2021
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Institucional da UFMG
• Texto Completo: http://hdl.handle.net/1843/39046; https://orcid.org/0000-0003-2332-4457
• topic: Kumaraswamy-G; survival analysis; regression models; KGsurv package.; Estatística – Teses; Análise de sobrevivência (Biometria) – Teses; Análise de regressão – Teses

In this work, a new package in language R (called KGsurv) was proposed using algorithms implemented in Stan software, to model new models based on the Kumaraswamy-G family of distributions. We present the Kumaraswamy-G regression models for the following classes: proportional hazards, proportional odds, and accelerated failure time considering Exponential, Weibull, Gamma, Log-logistics, and Log-normal distributions to model the G distribution. In this work, the frequentist approach was considered to estimate the parameters of the models with right-censored survival data under the assumption of a non-informative censoring mechanism. Finally, applications were presented using three real data sets widely used in the survival analysis literature to verify the results of the models implemented in the KGsurv package.