bm_CVnnet: Cross-validation for Neural Networks

In biomod2: Ensemble Platform for Species Distribution Modeling

Cross-validation for Neural Networks

Description

This internal biomod2 function allows the user to compute cross-validation for neural networks in ANN model (see nnet and BIOMOD_Modeling).

Usage

bm_CVnnet(
  Input,
  Target,
  size = c(2, 4, 6, 8),
  decay = c(0.001, 0.01, 0.05, 0.1),
  maxit = 200,
  nbCV = 5,
  weights = NULL,
  seedval = 555
)

Arguments

Input	complete dataset with explanatory variables
Target	calibration dataset with observed presence / absence
size	(see parameter ANN$size in BIOMOD_ModelingOptions)
decay	(see parameter ANN$decay in BIOMOD_ModelingOptions)
maxit	(see parameter ANN$maxit in BIOMOD_ModelingOptions)
nbCV	(see parameter ANN$nbCV in BIOMOD_ModelingOptions)
weights	a vector of numeric values corresponding to weights over calibration lines
seedval	an integer value corresponding to the new seed value to be set

Value

A data.frame containing the following elements :

• Size : the size

• Decay : the decay value

• AUC : the corresponding Area Under Curve

Author(s)

Damien Georges

See Also

nnet, auc, roc, BIOMOD_ModelingOptions, BIOMOD_Modeling, bm_SampleBinaryVector, bm_RunModelsLoop

Other Secundary functions: bm_BinaryTransformation(), bm_CrossValidation(), bm_FindOptimStat(), bm_MakeFormula(), bm_PlotEvalBoxplot(), bm_PlotEvalMean(), bm_PlotRangeSize(), bm_PlotResponseCurves(), bm_PlotVarImpBoxplot(), bm_PseudoAbsences(), bm_RunModelsLoop(), bm_SRE(), bm_SampleBinaryVector(), bm_SampleFactorLevels(), bm_VariablesImportance()

biomod2 documentation built on July 9, 2023, 6:05 p.m.