BayesianMixtureFit-methods: BayesianMixtureFit of the Event Data

In scientific-computing-solutions/eventTools: Additional Extensions/Functionality for the eventPrediction Package

Description

Creates an EventModel based on the fitting a Bayesian MCMC mixture model using run.jags and jags.

BayesianMixtureFit the Event Data

Usage

BayesianMixtureFit(object, prior.init, ...)

## S4 method for signature 'EventData,PriorAndInitsMixture'
BayesianMixtureFit(object,
  prior.init, dist = "weibull", N.samples = 2000, N.burnin = 1000,
  N.chains = 1, N.thin = 1, seed = NULL, parallel = FALSE,
  N.proc = NULL, ...)

Arguments

object	The EventData object
prior.init	Initial values to parameters and prior values, see PriorAndInitValues
...	Additional arguments to be passed to the method
dist	Distribution (only Weibull available for mixture). Can "force" into exponential by specifying a tight boundary around 1 for the shape parameter.
N.samples	The number of samples to take (2000). This will also determine the number of simulations used in the prediction step.
N.burnin	Number of burnin iterations before sampling starts (1000). Before these 1000 adaptive iterations are tun to enhance the sampling efficiency.
N.chains	Number of chins to use. IF running on a parallel machine, one may match this with the number of processors (N.proc).
N.thin	Thining interval to use. May reduce auto-correlation.
seed	Random seed, integer (NULL)
parallel	Use more than one processor (FALSE)
N.proc	Number of processors to use if running on a cluster.

Value

EventData object with the data censored at the appropriate point

scientific-computing-solutions/eventTools documentation built on May 29, 2019, 3:44 p.m.