The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]metanet.ch.

bean: an overview

bean reduces sampling bias in species occurrence data by thinning it in environmental space rather than in geographic space. The result is a cleaner training set for species distribution models (SDM / ENM).

The protocol is:

1. Prepare raw occurrences with prepare_bean().
2. Choose a grid resolution with find_env_resolution(), which selects a kernel-density bandwidth for each environmental variable.
3. Thin occurrences with thin_env_nd() (stochastic) or thin_env_center() (deterministic).
4. Fit a niche ellipsoid with fit_ellipsoid().
5. Predict suitability with predict() on the fitted ellipsoid.

library(bean)

data(origin_dat_prepared, package = "bean")
env_vars <- c("bio_1", "bio_4", "bio_12", "bio_15")

# 1. Pick an objective grid resolution from the data
res <- find_env_resolution(origin_dat_prepared, env_vars = env_vars)
res
#> --- Bean environmental grid resolution ---
#> Bandwidth selector: sheather-jones
#> 
#>  variable  resolution
#>     bio_1 0.056162684
#>     bio_4 0.013438324
#>    bio_12 0.004615848
#>    bio_15 0.006501067

# 2. Thin in environmental space
thinned <- thin_env_nd(
  data = origin_dat_prepared,
  env_vars = env_vars,
  grid_resolution = res$suggested_resolution,
  seed = 1
)
thinned
#> --- Bean Stochastic Thinning Results ---
#> 
#> Thinned 1024 original points to 78 points.
#> This represents a retention of 7.6% of the data.
#> 
#> --------------------------------------

These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.