run_optim: Run maximum likelihood optimisation of a model on simulated...

In TheoPannetier/DDvTDtools: Compare diversity-dependent (DD) and time-dependent (TD) diversification models

Description

Read a set of simulated trees as created by run_sim(), set up initial parameter values, call DDD::dd_ML() or DDD::bd_ML() to run the optimisation and format output.

Usage

run_optim(
  sim,
  optim,
  para,
  custom_pars = NULL,
  outputfile = paste0("sim", sim, "_optim", optim, "_", para, "_true_k.rds"),
  rangemc = 1:1000,
  methode = "ode45",
  optimmethod = "subplex",
  tol = rep(1e-06, 3),
  save_results = TRUE,
  return_results = FALSE,
  jobID = NA,
  num_cycles = 1,
  cond = 1
)

Arguments

sim	character, name of the simulation model, either "DD" or "TD".
optim	character, name of the optimisation model, either "DD" or "TD".
para	numeric, a four-digits number coding for a set of four parameter values. Refer to arg_para() doc for details, and call arg_para() to see possible inputs.
custom_pars	numeric vector, user-specified initial values for the parameters to be optimized. If NULL (default), values are initialized using true values used to simulate the trees.
outputfile	character, the name of the file to save the output data frame in. The default follows the structure expected by read_optim_results.
rangemc	numeric vector, a set of tree indices ranging from 1 to 1000.
methode	likelihood solving methode, passed to DDD::dd_loglik() / DDD::bd_loglik().
optimmethod	optimisation algorithm, passed to DDD::dd_ML() / DDD::bd_ML().
tol	optimisation tolerance, passed to DDD::dd_ML() / DDD::bd_ML().
save_results	logical. Should the results be saved to outputfile?
return_results	logical. Should the results be returned?
jobID	SLURM job ID passed when the function is called from a cluster script.
num_cycles	number of cycles of optimisation, passed to DDD::dd_ML() / DDD::bd_ML().
cond	code specifying how the likelihood is conditioned. Passed to DDD::dd_loglik() / DDD::bd_loglik().

Author(s)

Théo Pannetier

TheoPannetier/DDvTDtools documentation built on Oct. 22, 2020, 2:31 p.m.