gmjmcmc {FBMS}

Main algorithm for GMJMCMC (Genetically Modified MJMCMC)

Description

Main algorithm for GMJMCMC (Genetically Modified MJMCMC)

Usage

gmjmcmc(
  data,
  loglik.pi = gaussian.loglik,
  loglik.alpha = gaussian.loglik.alpha,
  transforms,
  P = 10,
  N.init = 100,
  N.final = 100,
  probs = NULL,
  params = NULL,
  sub = FALSE,
  verbose = TRUE
)

Arguments

data

A matrix containing the data to use in the algorithm, first column should be the dependent variable, second should be the intercept and the rest of the columns should be the independent variables.

loglik.pi

The (log) density to explore

loglik.alpha

The likelihood function to use for alpha calculation

transforms

A Character vector including the names of the non-linear functions to be used by the modification and the projection operator.

P

The number of generations for GMJMCMC (Genetically Modified MJMCMC). The default value is $P = 10$. A larger value like $P = 50$ might be more realistic for more complicated examples where one expects a lot of non-linear structures.

N.init

The number of iterations per population (total iterations = (T-1)*N.init+N.final)

N.final

The number of iterations for the final population (total iterations = (T-1)*N.init+N.final)

probs

A list of the various probability vectors to use

params

A list of the various parameters for all the parts of the algorithm

sub

An indicator that if the likelihood is inexact and should be improved each model visit (EXPERIMENTAL!)

verbose

A logical denoting if messages should be printed

Value

A list containing the following elements:

models

All models per population.

lo.models

All local optimization models per population.

populations

All features per population.

marg.probs

Marginal feature probabilities per population.

model.probs

Marginal feature probabilities per population.

model.probs.idx

Marginal feature probabilities per population.

best.margs

Best marginal model probability per population.

accept

Acceptance rate per population.

accept.tot

Overall acceptance rate.

best

Best marginal model probability throughout the run, represented as the maximum value in unlist(best.margs).

Examples

result <- gmjmcmc(matrix(rnorm(600), 100), P = 2, gaussian.loglik, NULL, c("p0", "exp_dbl"))
summary(result)
plot(result)