parse_lambdas: Parse Maxent lambdas information

View source: R/parse_lambdas.R

parse_lambdasR Documentation

Parse Maxent lambdas information

Description

Parse Maxent .lambdas files to extract the types, weights, minima and maxima of features, as well as the fitted model's entropy and other values required for predicting to new data.

Usage

parse_lambdas(lambdas)

Arguments

lambdas

Either a MaxEnt fitted model object (fitted with the maxent function in the dismo package), or a file path to a Maxent .lambdas file.

Value

A list (of class lambdas) with five elements:

  • lambdas: a data.frame describing the features used in a Maxent model, including their weights (lambdas), maxima, minima, and type;

  • linearPredictorNormalizer: a constant that ensures the linear predictor (the sum of clamped features multiplied by their respective feature weights) is always negative (for numerical stability);

  • densityNormalizer: a scaling constant that ensures Maxent's raw output sums to 1 over background points;

  • numBackgroundPoints: the number of background points used in model training; and

  • entropy: the entropy of the fitted model.

References

See Also

read_mxe() project()

Examples

# Below we use the dismo::maxent example to fit a Maxent model:
if (require(dismo) && require(rJava) && 
    file.exists(system.file('java/maxent.jar', package='dismo'))) {
  fnames <- list.files(system.file('ex', package="dismo"), '\\.grd$', 
                       full.names=TRUE )
  predictors <- stack(fnames)
  occurrence <- system.file('ex/bradypus.csv', package='dismo')
  occ <- read.table(occurrence, header=TRUE, sep=',')[,-1]
  me <- maxent(predictors, occ, path=file.path(tempdir(), 'example'), 
               factors='biome')

  # ... and then parse the lambdas information:
  lam <- parse_lambdas(me)
  lam
  str(lam, 1)
  
  parse_lambdas(file.path(tempdir(), 'example/species.lambdas'))
  
}

johnbaums/rmaxent documentation built on Oct. 11, 2024, 11:14 a.m.