estimate_migration_ABC: Estimates migration parameters using an ABC method.

In FlywayNet: Estimate Migration Birds Flyways Network with HSMM

Estimates migration parameters using an ABC method.

Description

Estimates migration parameters using an ABC (Approximate Bayesian Approach) method. Use the Lenormand method (ABC_sequential method) from the EasyABC package.

Usage

estimate_migration_ABC(
  migr,
  estimate_transitions = TRUE,
  estimate_sojourns = TRUE,
  sojourn_domain = NULL,
  nb_simul = 1000,
  only_likelihood = FALSE,
  p_acc_min = 0.05,
  choice_method = "mode",
  venter_bw = 0.2,
  n_cluster = 1,
  verbose = TRUE
)

Arguments

migr	A migration structure. Required migration fields are: site_name, link_knowledge, flight_duration, initial_state, horizon, death_probability, observations. Field transition_law_param is also required if the attribute estimate_transitions is set to FALSE. Field sojourn_law_param is also required if the attribute estimate_sojourns contains boolean values set to FALSE.
estimate_transitions	If TRUE, transitions probabilities are estimated. Default value is TRUE.
estimate_sojourns	Vector of boolean, identifies sites for which mean sojourn times must be estimated. Default value is TRUE transformed in TRUE for every site except for the last one (no need to estimate sojourn duration on the last site which is an arrival site).
sojourn_domain	Range (min and max) of the sojourn time parameters. Possible values: (i) NULL, all intervals are set to [0, migr$horizon], (ii) vector of 2 values min and max then all domains are [min, max] interval, (iii) list of interval for each site. Note that are taken into account only intervals for sites with TRUE in estimate_sojourn arguments. Default value is NULL.
nb_simul	Number of simulations divided by 2 to return in the posterior distribution (see Lenormand method). A positive integer. Default value is 1000.
only_likelihood	Boolean. If TRUE, the vector of statistics in ABC is composed of this log likelihood only. If FALSE, the vector of statistics is composed of the full matrix (site*time) of simulated observations. Default is FALSE.
p_acc_min	The Lenormand method parameter. A positive number between 0 and 1 (strictly). This is the stopping criterion of the algorithm: a small number ensures a better convergence of the algorithm, but at a cost in computing time. Default value is 0.05.
choice_method	Name of method (in "mode", "mean", "median","density") to choose parameters from ABC distribution. Default is "mode".
venter_bw	if choice_method is "mode", the bandwidth in [0, 1] to be used with the modeest::venter method. Default value is 0.2.
n_cluster	The number of cores used for simulation. Default value is 1.
verbose	If TRUE, display iterations achievement. Default is TRUE.

Value

The migration structure given with a new attribute estimation_method which is a structure with 3 attributes:

. name = "ABC",

. settings, a structure with attributes: estimate_transitions, estimate_sojourns, sojourn_domain,
nb_simul, only_likelihood, p_acc_min, choice_method, venter_bw, n_cluster,

. output, a structure with attributes:

- transition_law_param: estimated transition parameters,

- sojourn_law_param: estimated sojourn parameters,

- log_ABC: see EasyABC::ABC_sequential documentation for attributes param, stats, stats_normalization, weigths, epsilon, nsim, computime,

- log_param_dist: distribution of free estimated parameters (dataframe size nb_simul/2 x number of estimated free parameters),

- log_param_def: free parameters estimated with their initial values.

Examples

migr <- generate_toy_migration()
# Argument nb_simul just provide to reduce running time
estimated_migr <- estimate_migration_ABC( migr, nb_simul = 30 ) 
print( estimated_migr$estimation_method$output$transition_law_param )
estimated_migr <- estimate_migration_ABC( migr, nb_simul = 40 , sojourn_domain = c(2,4), 
                                         estimate_sojourns = c(TRUE, TRUE, TRUE, FALSE, FALSE))
print( estimated_migr$estimation_method$output$transition_law_param )

FlywayNet documentation built on March 18, 2022, 7:21 p.m.