eval_log_like_knobj: Posterior function.
In pauwels2014: Bayesian Experimental Design for Systems Biology.

Description Usage Arguments Details Value Author(s) See Also Examples

Computes the posterior value associated to a given parameter value for a given knowledge list.

1	eval_log_like_knobj(theta, knobj, fail_incoming = F, fit = F)

`theta`	A parameter named numeric vector.
`knobj`	A knowledge list. See `knobjs`
`fail_incoming`	A boolean indicating wether an error message is given by the ode solver.
`fit`	A parameter to be passed to the likelihood function. It indicates wether further prior information about the smoothness of the dynamical time course should be considered. This is used to guide local search posterior maximization methods.

The function computes the log prior first and then the likelihood associated to all the time course data found in the knobj$datas slot. The likelihood terms are summed. The prior term and the likelihood terms are weighted, weights being the inverse the number of observations they represent. This is necessary in order to give comparable contributions to low resolution and high resolution experiments.

A numerical value if fail_incoming == FALSE. A list containing a res numerical slot and a fail boolean slot representing wether the ode solver failed or not.

Edouard Pauwels

log_prior, knobjs, eval_kn_log_like

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
  ]
 )
)

eval_log_like_knobj(knobj$global_parameters$true_params_T, knobj)