ipmr: make_ipm – R documentation

Become an expert in R — Interactive courses, Cheat Sheets, certificates and more!

make_ipm

Methods to implement an IPM

Description

The make_ipm.* methods convert a proto_ipm into a set of discretized kernels and population vectors. Methods have different requirements, so be sure to read the parameter documentation. vignette('ipmr-introduction', 'ipmr') a more complete introduction.

Usage

make_ipm(
  proto_ipm,
  return_main_env = TRUE,
  return_all_envs = FALSE,
  usr_funs = list(),
  ...
)

## S3 method for class 'simple_di_det'
make_ipm(
  proto_ipm,
  return_main_env = TRUE,
  return_all_envs = FALSE,
  usr_funs = list(),
  ...,
  domain_list = NULL,
  iterate = TRUE,
  iterations = 50,
  normalize_pop_size = TRUE,
  iteration_direction = "right"
)

## S3 method for class 'simple_di_stoch_kern'
make_ipm(
  proto_ipm,
  return_main_env = TRUE,
  return_all_envs = FALSE,
  usr_funs = list(),
  ...,
  domain_list = NULL,
  iterate = TRUE,
  iterations = 50,
  kernel_seq = NULL,
  normalize_pop_size = TRUE,
  report_progress = FALSE,
  iteration_direction = "right"
)

## S3 method for class 'simple_di_stoch_param'
make_ipm(
  proto_ipm,
  return_main_env = TRUE,
  return_all_envs = FALSE,
  usr_funs = list(),
  ...,
  domain_list = NULL,
  iterate = TRUE,
  iterations = 50,
  kernel_seq = NULL,
  normalize_pop_size = TRUE,
  report_progress = FALSE,
  iteration_direction = "right"
)

## S3 method for class 'general_di_det'
make_ipm(
  proto_ipm,
  return_main_env = TRUE,
  return_all_envs = FALSE,
  usr_funs = list(),
  ...,
  domain_list = NULL,
  iterate = TRUE,
  iterations = 50,
  normalize_pop_size = TRUE,
  iteration_direction = "right"
)

## S3 method for class 'general_di_stoch_kern'
make_ipm(
  proto_ipm,
  return_main_env = TRUE,
  return_all_envs = FALSE,
  usr_funs = list(),
  ...,
  domain_list = NULL,
  iterate = TRUE,
  iterations = 50,
  kernel_seq = NULL,
  normalize_pop_size = TRUE,
  report_progress = FALSE,
  iteration_direction = "right"
)

## S3 method for class 'general_di_stoch_param'
make_ipm(
  proto_ipm,
  return_main_env = TRUE,
  return_all_envs = FALSE,
  usr_funs = list(),
  ...,
  domain_list = NULL,
  iterate = TRUE,
  iterations = 50,
  kernel_seq = NULL,
  normalize_pop_size = TRUE,
  report_progress = FALSE,
  iteration_direction = "right"
)

## S3 method for class 'simple_dd_det'
make_ipm(
  proto_ipm,
  return_main_env = TRUE,
  return_all_envs = FALSE,
  usr_funs = list(),
  ...,
  domain_list = NULL,
  iterate = TRUE,
  iterations = 50,
  normalize_pop_size = FALSE,
  report_progress = FALSE,
  iteration_direction = "right"
)

## S3 method for class 'simple_dd_stoch_kern'
make_ipm(
  proto_ipm,
  return_main_env = TRUE,
  return_all_envs = FALSE,
  usr_funs = list(),
  ...,
  domain_list = NULL,
  iterate = TRUE,
  iterations = 50,
  kernel_seq = NA_character_,
  normalize_pop_size = FALSE,
  report_progress = FALSE,
  iteration_direction = "right"
)

## S3 method for class 'simple_dd_stoch_param'
make_ipm(
  proto_ipm,
  return_main_env = TRUE,
  return_all_envs = FALSE,
  usr_funs = list(),
  ...,
  domain_list = NULL,
  iterate = TRUE,
  iterations = 50,
  kernel_seq = NA_character_,
  normalize_pop_size = FALSE,
  report_progress = FALSE,
  iteration_direction = "right"
)

## S3 method for class 'general_dd_det'
make_ipm(
  proto_ipm,
  return_main_env = TRUE,
  return_all_envs = FALSE,
  usr_funs = list(),
  ...,
  domain_list = NULL,
  iterate = TRUE,
  iterations = 50,
  normalize_pop_size = FALSE,
  report_progress = FALSE,
  iteration_direction = "right"
)

## S3 method for class 'general_dd_stoch_kern'
make_ipm(
  proto_ipm,
  return_main_env = TRUE,
  return_all_envs = FALSE,
  usr_funs = list(),
  ...,
  domain_list = NULL,
  iterate = TRUE,
  iterations = 50,
  kernel_seq = NA_character_,
  normalize_pop_size = FALSE,
  report_progress = FALSE,
  iteration_direction = "right"
)

## S3 method for class 'general_dd_stoch_param'
make_ipm(
  proto_ipm,
  return_main_env = TRUE,
  return_all_envs = FALSE,
  usr_funs = list(),
  ...,
  domain_list = NULL,
  iterate = TRUE,
  iterations = 50,
  kernel_seq = NA_character_,
  normalize_pop_size = FALSE,
  report_progress = FALSE,
  iteration_direction = "right"
)

Arguments

`proto_ipm`	A proto_ipm. This should be the output of `define_kernel`, or the `define_*` functions.
`return_main_env`	A logical indicating whether to return the main environment for the model. This environment contains the integration mesh, weights, and other potentially useful variables for subsequent analyses. Default is `TRUE`.
`return_all_envs`	A logical indicating whether to return the environments that the kernel expressions are evaluated in. These may be useful for some analyses, such as regression-level sensitivity/elasticity analyses, but can also rapidly increase memory consumption for models with many kernels (e.g. ones with hierarchical effects that have many levels, or any `*_stoch_param` model). Default is `FALSE`.
`usr_funs`	An optional list of user-specified functions that are passed on to the model building process. This can help make vital rate expressions more concise and expressive. Names in this list should exactly match the names of the function calls in the `...` or `formula`.
`...`	Other arguments passed to methods.
`domain_list`	An optional list of new domain information to implement the IPM with.
`iterate`	A logical indicating whether or not iterate the model before exiting or just return the sub-kernels. Only applies to density-independent, deterministic models and density-independent, stochastic kernel re-sampled models.
`iterations`	If `iterate` is `TRUE`, then the number of iterations to run the model for.
`normalize_pop_size`	A logical indicating whether to re-scale the population vector to sum to 1 before each iteration. Default is `TRUE` for `_di_` methods and `FALSE` for `_dd_` methods.
`iteration_direction`	Either `"right"` (default) or `"left"`. This controls the direction of projection. Right iteration will generate the right eigenvector (if it exists), while left iteration generates the left eigenvector. These correspond to the stable trait distributions, and reproductive values, respectively. This parameter is mostly used internally by other functions. Use with care.
`kernel_seq`	For `*_stoch_kern` methods, the sequence of kernels to use during the simulation process. It should have the same number of entries as the number of `iterations`. This should be a vector containing levels of the hierarchical effects specified in `levels_hier_effs`, or empty. Support for Markov chains will eventually get implemented. If it is empty, `make_ipm` will try to generate a sequence internally using a random selection of the `levels_hier_effs` defined in `define_kernel`.
`report_progress`	A logical indicating whether or not to periodically report progress for a stochastic simulation. Does not apply to deterministic methods. Default is `FALSE`.

Value

The make_ipm.* methods will always return a list of length 6 containing the following components:

sub_kernels: a list of arrays specified in define_kernel.
env_list: a list containing the evaluation environments of kernel. This will contain the main_env object if return_main_env = TRUE. It will also contain the sub-kernels evaluation environments if return_all_envs = TRUE.
env_seq: a character vector with length iterations of kernel indices indicating the order in which kernels are to be/were resampled OR a matrix with as many columns as stochastic parameters and n_iterations rows.
pop_state: population vectors stored as a list of arrays. The first dimension of each array corresponds to the state variable distribution, and the second dimension corresponds to time steps.
proto_ipm: the proto_ipm object used to implement the model.

In addition to the list class, each object will have a class comprised of the arguments from init_ipm plus 'ipm' pasted together with underscores. This is to facilitate print, plot, and lambda methods. For example, a init_ipm("general", "di", "det") model will have the class 'general_di_det_ipm' once it has been implemented using make_ipm.

ipmr

Fits Integral Projection Models Using an Expression Based Framework

v0.0.1

MIT + file LICENSE

Authors

Sam Levin [aut, cre] (<https://orcid.org/0000-0002-3289-9925>), Aldo Compagnoni [aut], Dylan Childs [aut], Sanne Evers [aut], Roberto Salguero-Gomez [aut], Tiffany Knight [aut]

Initial release