r - Error while using woe procedure in klaR package -


i have noticed reason procedure woe klar package produce error while trying execute on data.frame 1 column being one-level factor. here example:

a<-factor(rep(c("bad", "good"), 5)) b<-factor(rep(c(1,2), 5)) c<-factor(rep(c(2,3), 5)) d<-factor(rep(1,10)) df<-data.frame(a=a, b=b, c=c, d=d) 

now execution of

woe(a~., data = df[,-4]) 

won't produce errors, whereas

woe(a~., data = df) 

will crash following:

error in applywoes(object[[i]], x.fact[, which(names(x.fact) == names(object)[i])]) :    factor levels not match! in addition: warning message: in is.na(e2) : is.na() applied non-(list or vector) of type 'null' 

seems me problem d has 1 level factor. nonetheless, definition of woe should give value of 0 such factor...

is error in procedure implementation or (for strange reason) deliberate behavior?

my guess not deliberate behaviour. read on git( https://github.com/cran/klar/blob/master/r/woe.r ), fonctions such "woe.default" start couple of check of common error cases. error checked. should contact directly team behind package proper answer.


Comments

Popular posts from this blog

javascript - Any ideas when Firefox is likely to implement lengthAdjust and textLength? -

matlab - "Contour not rendered for non-finite ZData" -

delphi - Indy UDP Read Contents of Adata -