BEAST v2.5.0 Documentation: beast.math.distributions.MarkovChainDistribution

BEASTObject that performs calculations based on the State.
Probabilistic representation that can produce a log probability for instance for running an MCMC chain.
A class that produces a distribution chaining values in a parameter through the Gamma distribution. The value of a parameter is assumed to be Gamma distributed with mean as the previous value in the parameter. If useLogNormal is set, a log normal distribution is used instead of a Gamma. If a Jeffrey's prior is used, the first value is assumed to be distributed as 1/x, otherwise it is assumed to be uniform. Handy for population parameters.

Logable: yes, this can be used in a log.

Inputs:

jeffreys, reverse, uselog, shape, parameter, useLogNormal

 

jeffreys
type: java.lang.Boolean
use Jeffrey's prior (default false)
Optional input. Default: false

 

reverse
type: java.lang.Boolean
parameter in reverse (default false)
Optional input. Default: false

 

uselog
type: java.lang.Boolean
use logarithm of parameter values (default false)
Optional input. Default: false

 

shape
type: java.lang.Double
shape parameter of the Gamma distribution (default 1.0 = exponential distribution) or precision parameter if the log normal is used.
Optional input. Default: 1.0

 

parameter
type: beast.core.parameter.RealParameter
chain parameter to calculate distribution over
Required input

 

useLogNormal
type: java.lang.Boolean
use Log Normal distribution instead of Gamma (default false)
Optional input. Default: false