INLAcircular - Bayesian Joint Circular Regression Using 'INLA'

Provides a Bayesian regression framework for joint models with circular responses and covariates using integrated nested Laplace approximations, following Ye, Van Niekerk and Rue (2026) <doi:10.48550/arXiv.2602.08413>. The package supports link-adjusted von Mises likelihoods, shared latent predictors, structured latent effects, and optional LKJ-based multivariate covariance models. It also provides compiled circular densities and d/p/q/r penalized-complexity priors for von Mises, cardioid, and wrapped Cauchy concentration parameters as developed by Ye, Van Niekerk and Rue (2026) <doi:10.1177/1471082X261453677>.