par_zero_inflation {coala}

Zero Inflation for Parameters

Description

This adds a zero inflation to the distribution of a parameter for the different loci. When using this, each locus will be simulated with a parameter value of 0 with probability prob, or with parameter's original value in the remaining cases. are called separately for each locus, which can dramatically increase the time needed to simulate models with many loci.

Usage

par_zero_inflation(par, prob, random = TRUE)

Arguments

par

A parameter which will be set to 0 for part of the loci.

prob

The probability that the parameters value will be set to 0 for each locus if random is TRUE. Otherwise, it's the fixed fraction of loci which will have a parameter value of 0.

random

Whether the number of loci which are simulated with a value of 0 should be random (TRUE) or a fixed fraction (FALSE). See prob for details.

See Also

For parameters that are identical for all loci: parameter

Examples

# Simulate half of the loci with recomination and the other half without it:
model <- coal_model(4, 4) +
  feat_recombination(par_zero_inflation(par_named("rho"), .5, random = FALSE)) +
  sumstat_trees()
simulate(model, pars = c(rho = 1))

# Use a zero inflated gamma distribution:
model <- coal_model(4, 4) +
  feat_recombination(par_zero_inflation(par_variation(1, 10), .3)) +
  sumstat_trees()
simulate(model)