| loadings {diproperm} | R Documentation |
Returns the loadings of the binary linear classifier (e.g. DWD)
Description
Returns the variable indexes who had the highest loadings in the binary classification. Higher loading values indicate a variable's contribution toward the separation between the two classes. The loadings vector is a unit vector; thus the individual loadings range from -1 to 1, and the sum of the squares of the loadings equals one. The loadings direction vector points from the negative to positive class. Thus, positive entries correspond to variables that tend to be larger for the positive class.
Usage
loadings(dpp, loadnum = length(dpp$w))
Arguments
dpp |
A DiProPerm list. |
loadnum |
An integer indicating the number of variables to display. For example, if loadnum=5 then the indexes for the five variables who contributed most toward the separation of the two classes are displayed. The default is to print out all the loadings. |
Value
Returns the indexes and loadings for variables which contributed the most toward the separation of the binary classifier.
Author(s)
Andrew G. Allmon, J.S. Marron, Michael G. Hudgens
Examples
data(mushrooms)
X <- Matrix::t(mushrooms$X)
y <- mushrooms$y
dpp <- DiProPerm(X=X,y=y,B=10)
loadings(dpp,loadnum=3)