I'm running a predictive model and I have one possible predictive variable that is a factor and has more than 800 levels. I tried to reduce it running ctree in R (with the variable as the only predictive variable and the original class as class) and using the prediction to bin the variable and reduce the levels. But ctree takes too long (and never ends). What other method would you recommend?
Asked
Active
Viewed 136 times
3
-
How many observations do you have? – kjetil b halvorsen Jun 24 '16 at 17:25