Skip to contents

Fit and validate Quantile Regression Random Forests models

Source: R/fit_abund_qrf.R
fit_abund_qrf.Rd

Fit and validate Quantile Regression Random Forests models

Usage

fit_abund_qrf(
  data,
  response,
  predictors,
  predictors_f = NULL,
  fit_formula = NULL,
  partition,
  predict_part = FALSE,
  framework = "quantregForest",
  train_quantiles = c(0.5),
  eval_quantile = 0.5,
  mtry = length(c(predictors, predictors_f))/3,
  ntree = 2000,
  nodesize = 5,
  verbose = TRUE
)

Arguments

data

tibble or data.frame. Database with response, predictors, and partition values

response

character. Column name with species abundance.

predictors

character. Vector with the column names of quantitative predictor variables (i.e. continuous variables). Usage predictors = c("temp", "precipt", "sand")

predictors_f

character. Vector with the column names of qualitative predictor variables (i.e. ordinal or nominal variables type). Usage predictors_f = c("landform")

fit_formula

formula. A formula object with response and predictor variables (e.g. formula(abund ~ temp + precipt + sand + landform)). Note that the variables used here must be consistent with those used in response, predictors, and predictors_f arguments. Default NULL

partition

character. Column name with training and validation partition groups.

predict_part

logical. Save predicted abundance for testing data. Default is FALSE.

framework

character. Specifies the quantile regression framework to use. Either "quantregForest" (default) for traditional quantile regression forest from quantregForest package, or "grf" for generalized random forests from grf package.

train_quantiles

numeric vector. Quantiles to be estimated during model training. Default c(0.5) for median.

eval_quantile

numeric. Specific quantile to use for model evaluation metrics. Must be one of the values in train_quantiles. Default 0.5.

mtry

numeric. Number of variables randomly sampled as candidates at each split. Default (length(c(predictors, predictors_f))/3)

ntree

numeric. Number of trees to grow. This should not be set to too small a number, to ensure that every input row gets predicted at least a few times. Default 500

nodesize

numeric. Minimum size of terminal nodes. Controls tree depth - larger values produce smaller trees. Default 5 for quantregForest framework.

verbose

logical. If FALSE, disables all console messages. Default TRUE

Value

A list object with:

  • model: A "randomForest" class object from randomForest package. This object can be used for predicting.

  • predictors: A tibble with quantitative (c column names) and qualitative (f column names) variables use for modeling.

  • performance: Averaged performance metrics (see adm_eval).

  • performance_part: Performance metrics for each replica and partition.

  • predicted_part: Observed and predicted abundance for each test partition.

See also

fit_abund_cnn, fit_abund_dnn, fit_abund_gam, fit_abund_gbm, fit_abund_glm, fit_abund_net, fit_abund_svm, fit_abund_xgb

Examples

if (FALSE) { # \dontrun{
require(terra)
require(dplyr)

# Database with species abundance and x and y coordinates
data("sppabund")

# Extract data for a single species
some_sp <- sppabund %>%
  dplyr::filter(species == "Species one") %>%
  dplyr::select(-.part2, -.part3)

# Explore reponse variables
some_sp$ind_ha %>% range()
some_sp$ind_ha %>% hist()

# Here we balance number of absences
some_sp <-
  balance_dataset(some_sp, response = "ind_ha", absence_ratio = 0.2)
} # }