mahout-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andrew Musselman <andrew.mussel...@gmail.com>
Subject Random forest
Date Thu, 19 Sep 2013 16:47:29 GMT
Hi group, does the label/target variable need to be binary or can it be categorical with many
values?

Also, does the training set need to have equal numbers of each possible label value?

I'm finding that when I train on a binary label with many more of value b than value a I don't
get any predictions in the confusion matrix for a.  Is that a known issue?

Thanks
Andrew
Mime
  • Unnamed multipart/alternative (inline, 7-Bit, 0 bytes)
View raw message