find_best_model: Function to find the best n-dimensional ellipsoid model using...

In luismurao/hdm: hdm: "Habitat distribution models"

Description

Function to find the best n-dimensional ellipsoid model using Partial Roc as a performance criteria.

Usage

find_best_model(this.species, cor_threshold = 0.9, ellipsoid_level = 0.975,
  nvars_to_fit = 3, E = 0.05, RandomPercent = 50, NoOfIteration = 1000)

Arguments

this.species,	Species Temporal Environment "sp.temporal.env" object see extract_by_year.
cor_threshold	Threshold valuefrom which it is considered that the correlation is high see correlation_finder.
ellipsoid_level	The proportion of points to be included inside the ellipsoid see ellipsoidfit.
nvars_to_fit	Number of variables that will be used to model.
E	Amount of error admissible for Partial Roc test (by default =.05). Value should range between 0 - 1. see PartialROC
RandomPercent	Occurrence points to be sampled in randomly for the boostrap of the Partial Roc test PartialROC.
NoOfIteration	Number of iteration for the bootstrapping of the Partial Roc test PartialROC.

Value

A "sp.temp.best.model" object with metadata of the best model given the performance of the Partial Roc test.

luismurao/hdm documentation built on May 24, 2019, 5:06 a.m.