fscaret {fscaret}

feature selection caret

Description

Main function for fast feature selection. It utilizes other functions as regPredImp or impCalc to obtain results in a list of data frames.

Usage

fscaret(trainDF, testDF, installReqPckg = FALSE, preprocessData = FALSE,
	with.labels = TRUE, classPred = FALSE, regPred = TRUE, skel_outfile = NULL,
	impCalcMet = "RMSE&MSE", myTimeLimit = 24 * 60 * 60, Used.funcRegPred = NULL,
	Used.funcClassPred = NULL, no.cores = NULL, method = "boot", returnResamp = "all",
	missData=NULL, supress.output=FALSE, saveModel=FALSE, lvlScale=FALSE, ...)

Arguments

trainDF

Data frame of training data set, MISO (multiple input single output) type

testDF

Data frame of testing data set, MISO (multiple input single output) type

installReqPckg

If TRUE prior to calculations it installs all required packages, please be advised to be logged as root (admin) user

preprocessData

If TRUE data preprocessing is performed prior to modeling

with.labels

If TRUE header of the input files are read

classPred

If TRUE classification models are applied. Please be advised that importance is scaled according to F-measure regardless impCalcMet settings.

regPred

If TRUE regression models are applied

skel_outfile

Skeleton output file, e.g. skel_outfile=c("_myoutput_")

impCalcMet

Variable importance calculation scaling according to RMSE and MSE, for both please enter impCalcMet="RMSE&MSE"

myTimeLimit

Time limit in seconds for single model development

Used.funcRegPred

Vector of regression models to be used, for all available models please enter Used.funcRegPred="all"

Used.funcClassPred

Vector of classification models to be used, for all available models please enter Used.funcClassPred="all"

no.cores

Number of cores to be used for modeling, if NULL all available cores are used, should be numeric type or NULL

method

Method passed to fitControl of caret package

returnResamp

Returned resampling method passed to fitControl of caret package

missData

Handling of missing data values. Possible values: "delRow" - delete observations with missing values, "delCol" - delete attributes with missing values, "meanCol" - replace missing values with column mean.

supress.output

If TRUE output of modeling phase by caret functions are supressed. Only info which model is currently calculated and resulting variable importance.

saveModel

Logical value [TRUE/FALSE] if trained model should be embedded in final model.

lvlScale

Logical value [TRUE/FALSE] if additional scaling should be applied. For more information plase refer to impCalc().

...

Additional arguments, preferably passed to fitControl of caret package

Value

$ModelPred

List of outputs from caret model fitting

$VarImp

Data frames of variable importance and corresponding trained models

$PPlabels

Data frame of resulting preprocessed data set with original input numbers and names

$PPTrainDF

Training data set after preprocessing

$PPTestDF

Testing data set after preprocessing

$VarImp$model

Trained models

Note

Be advised when using fscaret function as it requires hard disk operations for saving fitted models and data frames. Files are written in R temp session folder, for more details see tempdir(), getwd() and setwd()

Author(s)

Jakub Szlek and Aleksander Mendyk

References

Kuhn M. (2008) Building Predictive Models in R Using the caret Package Journal of Statistical Software 28(5) http://www.jstatsoft.org/.

Examples

if((Sys.info()['sysname'])!="SunOS"){

library(fscaret)

# Load data sets
data(dataset.train)
data(dataset.test)

requiredPackages <- c("R.utils", "gsubfn", "ipred", "caret", "parallel", "MASS")

if(.Platform$OS.type=="windows"){

myFirstRES <- fscaret(dataset.train, dataset.test, installReqPckg=FALSE,
                  preprocessData=FALSE, with.labels=TRUE, classPred=FALSE,
                  regPred=TRUE, skel_outfile=NULL,
                  impCalcMet="RMSE&MSE", myTimeLimit=4,
                  Used.funcRegPred=c("lm"), Used.funcClassPred=NULL,
                  no.cores=1, method="boot", returnResamp="all",
                  supress.output=TRUE,saveModel=FALSE)

} else {

myCores <- 2

myFirstRES <- fscaret(dataset.train, dataset.test, installReqPckg=FALSE,
                  preprocessData=FALSE, with.labels=TRUE, classPred=FALSE,
                  regPred=TRUE, skel_outfile=NULL,
                  impCalcMet="RMSE&MSE", myTimeLimit=4,
                  Used.funcRegPred=c("lm","ppr"), Used.funcClassPred=NULL,
                  no.cores=myCores, method="boot", returnResamp="all",
                  supress.output=TRUE,saveModel=FALSE)

}



# Results
myFirstRES

}