Starting parameters for the optimization

Generate starting parameters for the optimization of the xsdm log-likelihood. Starting points are constructed from the environmental conditions at observed presences (where occ == 1) using a Latin hypercube design for the parameters based on the Sobol' low-discrepancy sequence.

Usage

start_parms(env_dat, mask = NULL, breadth = 1, num_starts = 100)

Arguments

env_dat: The environmental array for only the observed occurrences
mask: Either NULL or a named numeric vector. Names must be as specified by calling make_mask_names. The NULL case means planned optimizations will be over all model parameters, so start parameter sets should include all model parameters. The non-NULL case means some parameters will not be specified in the output of this function because the optimizations which are planned will fix those parameters anyway. The most common case for most applications will be mask=NULL.
breadth: Scalar in [0, 1] controlling how wide the search ranges are around their (fixed, data-driven) center. breadth = 1 (the default) reproduces the pre-v0.3 behaviour that corresponded to quant_vec = c(0.1, 0.5, 0.9); breadth = 0 collapses every range to essentially a single point, equivalent to quant_vec = c(0.5 - 1e-6, 0.5, 0.5 + 1e-6). Values in between interpolate linearly.
num_starts: The number of samples of the hypercube

Value

A data frame with samples for each parameter to optimize

Details

The bounds and center of the search range for a mu parameter are based on the quantiles in quant_vec applied to all observations of that environmental variable, over space and time. The relationship between quant_vec and the width of the ranges selected for the other parameters varies, but generally wider ranges in quant_vec produce wider ranges for start parameters.

Examples

env_dat <- example_1$env_array[example_1$occ_vec == 1, , ]
start_parms(env_dat)
#> # A tibble: 100 × 9
#>      mu1   mu2 sigltil1 sigltil2 sigrtil1 sigrtil2   ctil     pd o_par1
#>    <dbl> <dbl>    <dbl>    <dbl>    <dbl>    <dbl>  <dbl>  <dbl>  <dbl>
#>  1  9.58  4.00   0.364    0.359   0.172     0.620  -1.26   0.858 -4.86 
#>  2 11.8   5.29   1.06    -0.334   0.865    -0.0736 -0.624 -1.34   4.57 
#>  3 12.9   3.36   0.711    0.0127  1.21      0.273  -0.944  1.96  -0.147
#>  4 10.7   4.65   0.0178  -0.680   0.518    -0.420  -1.58  -0.240  9.28 
#>  5 11.2   3.04   0.884    0.533   1.04     -0.247  -1.42  -1.89  -7.22 
#>  6 13.4   4.33   0.191   -0.161   0.345     0.446  -0.784  0.309  2.21 
#>  7 12.3   3.68  -0.155    0.186  -0.00144   0.0997 -1.10  -0.790 -2.50 
#>  8 10.1   4.97   0.538   -0.507   0.692     0.793  -1.74   1.41   6.92 
#>  9 10.4   3.52  -0.0688  -0.0740  0.778     0.706  -1.18  -1.06  -8.39 
#> 10 12.6   4.81   0.624    0.619   0.0852    0.0130 -0.544  1.13   1.03 
#> # ℹ 90 more rows
start_parms(env_dat, mask = c(mu2 = 5, pd = 1))
#> # A tibble: 100 × 7
#>      mu1 sigltil1 sigltil2 sigrtil1 sigrtil2   ctil o_par1
#>    <dbl>    <dbl>    <dbl>    <dbl>    <dbl>  <dbl>  <dbl>
#>  1  9.58   0.299   -0.139    0.995   -0.290  -0.664 -1.91 
#>  2 11.8    0.993    0.554    0.302    0.403  -1.30   7.51 
#>  3 12.9   -0.0472   0.208    0.648    0.750  -0.984  2.80 
#>  4 10.7    0.646   -0.486   -0.0448   0.0564 -1.62  -6.63 
#>  5 11.2   -0.220    0.381    1.17     0.576  -1.46  -4.27 
#>  6 13.4    0.473   -0.312    0.475   -0.117  -0.824  5.15 
#>  7 12.3    0.126   -0.659    0.822   -0.463  -1.14   0.442
#>  8 10.1    0.819    0.0343   0.129    0.230  -0.504 -8.98 
#>  9 10.4    0.0395  -0.572    0.562    0.316  -0.584 -0.736
#> 10 12.6    0.733    0.121    1.25    -0.377  -1.22   8.69 
#> # ℹ 90 more rows