An overview on parametric quantile regression models and their computational implementation with applications to...

Area of Science:

• Statistics
• Biostatistics
• Computational Statistics

Background:

• Quantile regression estimates covariate relationships with any response quantile, offering an alternative to mean-based analysis.
• Recent advancements include statistical distributions for parametric quantile modeling.
• Parametric quantile regression provides a flexible framework for analyzing complex data distributions.

Purpose of the Study:

• To present an up-to-date overview of parametric quantile regression.
• To introduce a new computational R package for parametric quantile modeling.
• To demonstrate applications using biomedical and COVID-19 datasets.

Main Methods:

• Formulated fully parametric quantile regression by parameterizing baseline distributions.
• Introduced regression-based functional forms using link functions.
• Implemented an R package for estimation, model checking, and generating functions for 23 distributions.

Main Results:

• Reviewed a wide range of parametric quantile regression models.
• Developed an R package enabling the fitting of diverse distributions.
• Applied parametric and non-parametric models to biomedical and COVID-19 data, comparing their performance.

Conclusions:

• Parametric quantile regression offers a valuable alternative to distribution-free methods.
• The developed R package simplifies the application of various parametric quantile models.
• The study highlights the utility of parametric quantile regression in real-world biomedical and public health scenarios.