Bayesian quantile regression

Bayesian quantile regression

Internship Description

This project is about Bayesian quantile regression, where we make use of the parametric form of the likelihood to define parametric quantile regression (also for some discrete response distributions). The advantage with this model-based approach, are 1) that the posterior distribution, and quantities derived from it, are still valid, and 2) we can easily extend the scope of the R-INLA software (www.r-inla.org) to do the approximate Bayesian inference when we have a latent Gaussian model.​

Deliverables/Expectations

The students are expected to create Rmarkdown tutorials with (also new) case studies, and assist in extending quantile regression to other families within the R-INLA framework.

Faculty Name

Haavard Rue

Field of Study

​Statistics