estimate_parameters_sc_parallel: Estimate parameters of the self-correcting model using...

In ldmppr: Estimate and Simulate from Location Dependent Marked Point Processes

Estimate parameters of the self-correcting model using log-likelihood maximization in parallel

Description

Estimate the parameters of the self-correcting model using [nloptr::nloptr()] given a set of delta values. Makes use of parallel computation to allow the user to identify the optimal delta value more quickly given available computational resources.

Usage

estimate_parameters_sc_parallel(
  data,
  x_grid = NULL,
  y_grid = NULL,
  t_grid = NULL,
  delta_values = NULL,
  parameter_inits = NULL,
  upper_bounds = NULL,
  opt_algorithm = "NLOPT_LN_SBPLX",
  nloptr_options = list(maxeval = 400, xtol_rel = 1e-05, maxtime = NULL),
  verbose = TRUE,
  num_cores = parallel::detectCores() - 1,
  set_future_plan = FALSE
)

Arguments

data	a data frame of locations and sizes labelled ("x", "y", "size").
x_grid	a vector of grid values for x.
y_grid	a vector of grid values for y.
t_grid	a vector of grid values for t.
delta_values	a vector of delta values.
parameter_inits	a vector of parameter initialization values.
upper_bounds	a vector of upper bounds for time, x, and y.
opt_algorithm	the NLopt algorithm to use for maximization.
nloptr_options	a list of named options for nloptr including "maxeval", "xtol_rel", and "maxtime".
verbose	'TRUE' or 'FALSE' indicating whether to show progress of optimization.
num_cores	number of cores to use in parallel estimation.
set_future_plan	'TRUE' or 'FALSE' indicating whether to change the parallelization plan for use in the function.

Details

Details regarding the self-correcting model and the estimation procedure can be found in the references.

Value

a list containing the optimal obtained values and the full [nloptr::nloptr()] results for all provided delta values.

References

Møller, J., Ghorbani, M., & Rubak, E. (2016). Mechanistic spatio-temporal point process models for marked point processes, with a view to forest stand data. Biometrics, 72(3), 687–696. \Sexpr[results=rd]{tools:::Rd_expr_doi("10.1111/biom.12466")}.

Examples

# Note: This function is designed to be run in parallel and may be computationally intensive.


# Load the small example data
data(small_example_data)

# Define the grid values
x_grid <- seq(0, 25, length.out = 5)
y_grid <- seq(0, 25, length.out = 5)
t_grid <- seq(0, 1, length.out = 5)

# Define the delta values
delta_values <- seq(0.25, 1, by = 0.25)

# Define the parameter initialization values
parameter_inits <- c(1.5, 8.5, .015, 1.5, 3.2, .75, 3, .08)

# Define the upper bounds
upper_bounds <- c(1, 25, 25)

# Estimate the parameters in parallel
estimate_parameters_sc_parallel(
  data = small_example_data,
  x_grid = x_grid,
  y_grid = y_grid,
  t_grid = t_grid,
  delta_values = delta_values,
  parameter_inits = parameter_inits,
  upper_bounds = upper_bounds,
  opt_algorithm = "NLOPT_LN_SBPLX",
  nloptr_options = list(
    maxeval = 50,
    xtol_rel = 1e-2
  ),
  verbose = TRUE,
  set_future_plan = TRUE
)

ldmppr documentation built on April 4, 2025, 12:45 a.m.