brms

We have hosted the application brms in order to run this application in our online workstations with Wine or directly.

Quick description about brms:

brms is an R package by Paul Bürkner which provides a high-level interface for fitting Bayesian multilevel (i.e. mixed effects) models, generalized linear / non-linear / multivariate models using Stan as the backend. It allows R users to specify complex Bayesian models using formula syntax similar to lme4 but with far more flexibility (distributions, link functions, hierarchical structure, nonlinear terms, etc.). It supports model diagnostics, posterior predictive checking, model comparison, custom priors, and advanced features such as distributional regression.

Features:

• Supports a very wide range of distributions and link functions: linear, robust linear, count, ordinal, survival, response times, zero-inflated, hurdle, mixture models etc.
• Multilevel / hierarchical models: grouping factors, random effects etc.
• Nonlinear and smooth terms (nonlinear modeling) and autocorrelation / time series structures etcetera
• Flexible prior specification for all parts of the model (including custom priors)
• Model diagnostics and model comparison: posterior predictive checks, leave-one-out cross-validation (LOO), comparison of fits etc.
• Ability to work with user-specified covariance structures, meta-analytic standard errors, censoring/truncation etc.

Programming Language: R.
Categories: