fit_all_dsm: Fit all possible model with the covariates given.

In MathieuGenu/geffaeR: Generic Function for Abundance Estimation with R

Fit all possible model with the covariates given.

Description

Fit all possible model with the covariates given.

Usage

fit_all_dsm(
  distFit = NULL,
  segdata_obs,
  obsdata,
  response = "ind",
  predictors,
  likelihood = "negbin",
  esw = NULL,
  max_cor = 0.5,
  nb_max_pred = 3,
  complexity = 4,
  smooth_xy = TRUE,
  k = 5,
  splines_by = NULL,
  weighted = FALSE,
  random = NULL,
  soap = list(xt = NULL, knots = NULL),
  use_loo = FALSE
)

Arguments

distFit	distfit object. One of plot_detection output or directly an output of ds
segdata_obs	segdata data.frame with observation added with ajout_obs.
obsdata	Observation data.frame prepared with prepare_data_obs.
response	Response variable to choose between "ind" number of indiviudals or "obs" number of observation.
predictors	vector containing all covariates as character string.
likelihood	Likelihood must be one of "negbin", "poisson" or "tweedie". Default is "negbin".
esw	Value of effective-strip width.
max_cor	Maximum correlation threshold between two covariates tolerated.
nb_max_pred	Maximum number of covariates in a model.
complexity	k argument in smoothers.
smooth_xy	Controls the intercept to include or not a bivariate smooth on x and y : for prediction inside the prospected polygon, should be set to TRUE to obtain stable estimates for prediction outside, MUST be set to FALSE to keep extrapolation under control.
k	Number of models to return for inference.
splines_by	Interaction with splines given by one variable of segdata.
weighted	Either to compute weights or no.
random	variable of segdata_obs to include random effect.
soap	list to pass in order to use a soap-film smooth: must be prepared outside this function.
use_loo	Model selection with leave-one-out cross validation if TRUE, if FALSE it uses AIC to select models.

Value

This function return a list containing:

1. all_fits_binded : data.frame Containing all fitting infos on models.

2. best_models : list k best models with standardized value.

3. best_models4plotting : list k best models with raw values (for plotting).

Note

By default, use cubic B-splines with shrinkage (bs = 'cs').

MathieuGenu/geffaeR documentation built on March 23, 2022, 7:50 p.m.