Description Usage Arguments Value
The proposed burn-in is calculated by dividing the Markov chains into m
blocks and calculate the 95% credibility intervals and the respective mean.
Starting in the first one, the block is taken as burn-in as long as the mean of the current block is not in the credibility interval of the following block or vice versa.
The thinning rate is proposed by the first lag which leads to a chain autocorrelation less than dependence
.
It is not easy to automate these choices, so it is highly recommended to verify the chains manually.
1 | diagnostic(chain, dependence = 0.8, m = 10)
|
chain |
vector of Markov chain samples |
dependence |
allowed dependence for the chain |
m |
number of blocks |
vector of burn-in and thinning
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.