CurrentPred: Produce binary and filtered data within Biomod
In BIOMOD: Ensemble platform for species distribution modeling

Description Usage Arguments Details Value Author(s) See Also Examples

This function enables to create binary and/or filtered datasets from the probability data and the thresholds produced by the Models() function of Biomod for current data.

1 2	CurrentPred(GLM=TRUE, GBM=TRUE, GAM=TRUE, CTA=TRUE, ANN=TRUE, SRE=TRUE, FDA=TRUE, MARS=TRUE, RF=TRUE, BinRoc=FALSE, BinKappa=FALSE, BinTSS=FALSE, FiltRoc=FALSE, FiltKappa=FALSE, FiltTSS=FALSE)

`GLM`	type True or False to produce binary and/or filtered prediction data for that model
`GAM`	type True or False to produce binary and/or filtered prediction data for that model
`GBM`	type True or False to produce binary and/or filtered prediction data for that model
`CTA`	type True or False to produce binary and/or filtered prediction data for that model
`RF`	type True or False to produce binary and/or filtered prediction data for that model
`FDA`	type True or False to produce binary and/or filtered prediction data for that model
`SRE`	type True or False to produce binary and/or filtered prediction data for that model
`MARS`	type True or False to produce binary and/or filtered prediction data for that model
`ANN`	type True or False to produce binary and/or filtered prediction data for that model
`BinRoc`	set to True if you want the predictions converted from probabilities to binary data using Roc
`BinKappa`	set to True if you want the predictions converted from probabilities to binary data using Kappa
`BinTSS`	set to True if you want the predictions converted from probabilities to binary data using TSS
`FiltRoc`	set to True if you want the predictions converted from probabilities to filtered data using Roc
`FiltKappa`	set to True if you want the predictions converted from probabilities to filtered data using Kappa
`FiltTSS`	set to True if you want the predictions converted from probabilities to filtered data using TSS

The thresholds used are those stored in the Evaluation.results.Roc/Kappa/TSS objects produced by the Models() function. As a general reminder, do think about reloading these objects stored in the .RData file also produced by the Models() function from one R session to another .

No values are returned but a series of objects are produced in the "pred" folder (see Models for further details). Like for the predictions, the files are arrays of 4 diensions and produced per species. They contain the predictions in binary or filtered predictions data. See the structure in the examples below.

Wilfried Thuiller, Bruno Lafourcade

Models, PredictionBestModel

## Not run: 
data(Sp.Env)
data(CoorXY)

#This command is necessary for the run of BIOMOD as a new dataframe is produced for the Models function
Initial.State(Response=Sp.Env$Sp277, Explanatory=Sp.Env[,4:10], sp.name='Sp277',
IndependentResponse=NULL, IndependentExplanatory=NULL)

#Here are done a Full Calib. This will hence take several minutes.
Models(RF=TRUE, SRE=TRUE, GLM = FALSE, TypeGLM = "quad", Test = "AIC", CTA = FALSE, CV.tree = 50, ANN = FALSE, CV.ann = 2,
   NbRunEval = 1, DataSplit = 100, Roc=TRUE, Optimized.Threshold.Roc=TRUE, Kappa=TRUE, TSS=TRUE, VarImport=5,
   NbRepPA=0, strategy="circles", coor=CoorXY, distance=2, nb.absences=1000)


#The results of the Models() function are in probabilities (scaled between 0 to 1000). The CurrentPred() function
#enables to produce binary or filtered results. Here we decide to create binary data using the Roc and TSS thresholds, 
#and filtered data using only the TSS threshold.
CurrentPred(BinRoc=TRUE, BinKappa=FALSE, BinTSS=TRUE, FiltRoc=FALSE, FiltKappa=FALSE, FiltTSS=TRUE) 


#let's view some examples of them
load("pred/Pred_Sp277_BinRoc")
dim(Pred_Sp277_BinRoc)
dimnames(Pred_Sp277_BinRoc)[-1]
Pred_Sp277_BinRoc[120:140,,1,1]

load("pred/Pred_Sp277_FiltTSS")
Pred_Sp277_FiltTSS[120:140,,1,1]

level.plot(Pred_Sp277_FiltTSS[,"RF",1,1], CoorXY, show.scale=FALSE)

## End(Not run)