RSSAmplifier

Blog

Strong Inference

stronginference.comRSS feed ↗5 posts

Latest posts

The First Release of PyMC3

On Monday morning the PyMC dev team pushed the first release of PyMC3 , the culmination of over 5 years of collaborative work. We are very pleased to be able to provide a stable version of the package to the Python scientific computing community. For those of you unfamiliar with the …

Calculating Bayes factors with PyMC

Statisticians are sometimes interested in comparing two (or more) models, with respect to their relative support by a particular dataset. This may be in order to select the best model to use for inference, or to weight models so that they can be averaged for use in multimodel inference. The …

Burn-in, and Other MCMC Folklore

I have been slowly working my way through The Handbook of Markov Chain Monte Carlo , a compiled volume edited by Steve Brooks et al. that I picked up at last week's Joint Statistical Meetings. The first chapter is a primer on MCMC by Charles Geyer , in which he summarizes the …

Implementing Dirichlet processes for Bayesian semi-parametric models

Semi-parametric methods have been preferred for a long time in survival analysis, for example, where the baseline hazard function is expressed non-parametrically to avoid assumptions regarding its form. Meanwhile, the use of non-parametric methods in Bayesian statistics is increasing. However, there are few resources to …

Automatic Missing Data Imputation with PyMC

A distinct advantage of using Bayesian inference is in its universal application of probability models for providing inference. As such, all components of a Bayesian model are specified using probability distributions for either describing a sampling model (in the case of observed data) or characterizing the uncertainty of an unknown …