Potential Fix For VarImp To Work With Categorical Models Such as Method pda

[thomashigginsdatascientist]

insert on line 134 filterVarImp:

y_unlist = unlist(y)

change line 135 filterVarImp:

outStat <- apply(x, 2, testFunc, y = y_unlist)

insert on line 23 of varImp.train:

y_numeric <- is.numeric(y_dat)
if(y_numeric){
}else{
y_dat$.outcome <- as.factor(y_dat$.outcome)
y_dat$.outcome <- as.numeric(y_dat$.outcome)
}