Determine a starting value for ABC's delta

In ABC settings a model solution is compared to data with an acceptance threshold: delta. This-pre calibration function attempts to adjust this delta value.

Usage

preCalibration(
  objectiveFunction,
  npc = 1000,
  rprior,
  rep = 1,
  p = 0.05,
  sfactor = 0.1,
  delta = 0.01,
  num = 1
)

Arguments

objectiveFunction

function that, given a (vectorial) parameter as input, (1) simulates the model with the given parameter, and (2) outputs the distance between experimental data and simulated data simulated.

npc

sample size of pre-calibration.

rprior

a function that generates random ABC variables, distributed according to the prior.

rep

number of repetitions of the preCalibration process.

p

fraction (top scoring) of sampled points to base Sigma on (Sigma is the covariance matrix for the moves proposed in the ABCMCMC algorithm).

sfactor

scales Sigma up or down (Sigma is the covariance matrix for the moves proposed in the ABCMCMC algorithm).

delta

ABC threshold.

num

number of different starting parameter vectors (initial states of the chains) to generate. Usually, num is equal to the number of chain that will be run in the sampling procedure.

Value

list with entries prePar (sampled parameters), preDelta (distances between experimental data and trajectories produced with each of the parameters in prePar), Sigma (covariance matrix for the moves proposed in the ABCMCMC algortihm) and startPar (starting parameters for the ABCMCMC chains)