Название: Copula Additive Distributional Regression Using R
Автор: Giampiero Marra, Rosalba Radice
Издательство: CRC Press
Серия: The R Series
Год: 2026
Страниц: 151
Язык: английский
Формат: epub (true)
Размер: 10.1 MB

Copula additive distributional regression enables the joint modeling of multiple outcomes, an essential aspect of many real-world research problems. This book provides an accessible overview of this modeling approach, with a particular focus on its implementation in the GJRM R package, developed by the authors. The emphasis is on bivariate responses with empirical illustrations drawn from diverse fields such as health and medicine, epidemiology, economics and social sciences.

The book is organized into four parts. Part I introduces the copula additive distributional regression framework and provides a succinct overview of the GJRM and GJRM.data packages in R. Part II focuses on scenarios where the marginal distributions are of the same type, while Part III explores cases with mixed marginal types. Finally, Part IV demonstrates how copula regression can be applied to estimate causal treatment effects in the presence of unobserved confounding. The chapters in Parts II, III and IV follow a consistent structure: each begins with a description of a dataset that motivates the need for a particular copula model, determined primarily by the nature of the response variables of interest. This is followed by a discussion of the model and its log-likelihood, a description of the measure of interest for the chapters in Part IV, and a model fitting exercise using GJRM. Additional R code is available on the authors' websites. Readers are encouraged to start with Part I to grasp the fundamental concepts before progressing to the subsequent chapters. While some may choose to explore only those cases that align with their specific interests, reading the book comprehensively will provide a deeper understanding of the nuances of copula regression modeling.

Key Features:

Provides a comprehensive overview of joint regression modeling for multiple outcomes, with a focus on bivariate responses
Offers a practical approach with real-world examples from various fields
Demonstrates the implementation of all the discussed models using the GJRM package in R
Includes supplementary resources such as data accessible through the GJRM.data package in R and additional code available on the authors' webpages

This book is designed for graduate students, researchers, practitioners and analysts who are interested in using copula additive distributional regression for the joint modeling of bivariate outcomes. The methodology is accessible to readers with a basic understanding of core statistics and probability, regression, copula modeling and R.