The goal of envModel is to help prepare data and run models (currently only random forest) to predict group membership as a function of environmental variables. Reduce correlated and ‘unimportant’ environmental variables, run quick random forest runs over many clusterings outputting only confusion matrix, then choose an area of clustering space for a ‘good’ model and run random forest, iteratively adding trees until results stabilise between iterations.
Installation
envModel is not on CRAN.
You can install the development version from GitHub with:
# install.packages("devtools")
devtools::install_github("dew-landscapes/envModel")Contents of envModel
The following functions and data sets are provided in envModel.
| object | class | description |
|---|---|---|
envModel::get_conf_metrics() |
function | Calculate diagnostics |
envModel::make_env_clust_df() |
function | Add explanatory (environmental) variables to ‘site’ dataframe. |
envModel::make_env_corr() |
function | Generate correlation matrix and select variables to remove |
envModel::make_rf_diagnostics() |
function | Run random forest, returning only diagnostic values. |
envModel::make_rf_diagnostics_old() |
function | Run random forest, usually returning only diagnostic values. |
envModel::make_rf_diagnostics_spatialcv() |
function | Run random forest, returning only diagnostic values from spatial cross |
envModel::make_rf_good() |
function | Iteratively add trees to random forest until predictions stabilise |
envModel::reduce_env() |
function | Reduce number of environmental variables |