This function allow the user to select wanted descriptors between both that are intercorrelated with a correlation coefficent higher that ThresholdInterCor. The selection can also be automatic based on the correlation with the property of each variables.
Usage
select_variables(id, y, d, ThresholdInterCor, auto = FALSE)
Arguments
id
List of the names of observations
y
List of the values of the property/response
d
Dataframe containing the names of desciptors and their values (without missing or constant values)
ThresholdInterCor
Threshold value (double) of the accepted intercorrelation between descriptors (should be between 0 and 1)
auto
Two possible values: TRUE or FALSE (by default). The selection of descriptors is done automatically based on the correlation between descriptor and property (auto=TRUE) or is done manually by user (auto=FALSE)
Value
return a dataframe containing only of non intercorrelated variables
Examples
# Run after Preselection : d<-Preselection(desc)
# desc<-select_variables(id,y,d,0.95)