pauwels2014 (version 1.0)

sample_function_multi_mod_weight: Sample function visiting multiple modes of the posterior

Description

Generate a posterior sample using multiple local search maximization and sampling based on different initializations.

Usage

sample_function_multi_mod_weight(knobj)

Arguments

knobj

A knowledge list. See knobjs.

Value

A matrix which rows represent a named numeric vector of parameters

Details

The parameters governing the local search and sampling behaviour are defined in the global_parameters slot of the knobjs argument. The function consists in using the BFGS_special function to find an initialization for the Metropolis Hasting algorithm implented by generate_sample. This is done multiple times. This procedure is also applied to previous sample points. All those sample are aggregated and the resulting sample is chosen randomly based on the associated posterior values.

See Also

sample_function, BFGS_special, generate_sample

Examples

data(experiment_list1)
data(observables)

## Generate the knowledge object with correct parameter value
knobj <- generate_our_knowledge(transform_params)

## Initialize with some data
knobj$datas[[1]] <- list(
 manip = experiment_list1$nothing,
 data = add_noise(
  simulate_experiment(knobj$global_parameters$true_params_T, knobj, experiment_list1$nothing)[
   knobj$global_parameters$tspan %in% observables[["mrnaLow"]]$reso, 
   observables[["mrnaLow"]]$obs
  ]
 )
)

## Decrease parameter values for the example
knobj$global_parameters$max_it <- 2
knobj$global_parameters$n_multi_mod <- 2
knobj$global_parameters$sample_burn_in <- 5
knobj$global_parameters$sample_to_keep1 <- 100
knobj$global_parameters$final_sample <- 100
knobj$global_parameters$final_sample_est <- 100


#thetas <- sample_function_multi_mod_weight(knobj)
#thetas