MCMCResult
MCMCResult(chain: MarkovChain)
An object containing a constructed Markov chain.
Attributes
xs : Tensor
-
An \(n \times k\) matrix containing the samples that form the Markov chain.
potentials : Tensor
-
An \(n\)-dimensional vector containing the potential function associated with the target density evaluated at each sample in the chain.
acceptance_rate : float
-
The acceptance rate of the sampler.
iacts : Tensor
-
A \(k\)-dimensional vector containing estimates of the integrated autocorrelation time (IACT) for each parameter.
ess : Tensor
-
A \(k\)-dimensional vector containing estimates of the effective sample size (ESS) of each parameter.
Notes
The IACT for a given parameter is estimated according to \[ \hat{\tau}(M) = 1 + 2 \sum_{i=1}^{M} \hat{\rho}(i), \] where \(\hat{\rho}(i)\) denotes an estimate of the normalised autocorrelation function of the parameter for a lag of \(i\), \(M\) is the smallest integer such that \(M \geq C\hat{\tau}(M)\), and \(C=5\). For further information, see the emcee docs.
The ESS for a given parameter is estimated according to \[ N_{\mathrm{eff}} = \frac{1}{\hat{\tau}(M)}. \]