The SALTSampler package facilitates Monte Carlo Markov Chain (MCMC) sampling of random variables on a simplex. A Self-Adjusting Logit Transform (SALT) proposal is used so that sampling is still efficient even in difficult cases, such as those in high dimensions or with parameters that differ by orders of magnitude. Special care is also taken to maintain accuracy even when some coordinates approach 0 or 1 numerically. Diagnostic and graphic functions are included in the package, enabling easy assessment of the convergence and mixing of the chain within the constrained space.
|Author||Hannah Director, Scott Vander Wiel, James Gattiker|
|Date of publication||2015-11-03 01:13:45|
|Maintainer||Scott Vander Wiel <email@example.com>|
|License||BSD_3_clause + file LICENSE|
Diagnostics: Plots and Summaries of RunMh Output
GenData: Synthetic Data From a Multinomial Distribution
Logit: Logit of a Probability Vector
LogitScale: Finds logit(sp)
LogitSum: Log of the Sum of Probabilities
LogPq: Computes log(p) and log(1-p)
PropStep: Draw a Proposal on a Simplex
RunMh: Metropolis Hasting Algorithm Constrained on a Simplex
SALTSampler-package: Efficient Sampling on the Simplex
TriPlot: Plots MCMC Samples on a 3-Simplex
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.