All features significatly associated to the outcome will be decorrelated
featureDecorrelation(data=NULL,Outcome=NULL,
refdata=NULL,baseFeatures=NULL,loops=20,
thr=0.80,unipvalue=0.05,useDeCorr=TRUE,...)
predictDecorrelate(decorrelatedobject,testData)The dataframe whose features will de decorrelated
The target outcome
The frame used to get the correlation and lm formula
A vector of features to be used as the base vectors.
the maxumum number of loops
correlation threshold for feature selection
association p-value
if TRUE, the returned matrix will use the estimated decorrelation Matrix
parameters passed to the adjuting method.
The dataframe created by the featureDecorrelation function
Test data to decorrelate
The decorrelated data frame with the follwing attributes
The list of features that were decorrelated
The count of how many times was decorrelted
The Decorrelation matrix for linear models
The list of variables in the whitening matrix
The list of feetures used as base features
If TRUE the estimated DeCorrmatrix matrix was used for decorrelation
The data-frame will be analyzed and correlated features will be decorrelated. Feature selection may be based on Outcome or by default in Top correlation
featureAdjustment