Hi,

just was working with this and found this . The javadoc says

 * Encodes a categorical values with an unbounded vocabulary.  Values
are encoding by incrementing a
 * few locations in the output vector with a weight that is either
defaulted to 1 or that is looked
 * up in a weight dictionary.  By default, _only one probe is used_

which i immediately suspected was not true cause it doesn't make a lot
of default sense for unbounded categories.

Indeed, checking the code seems to indicate that by default 2 probes
are used for unbounded category encoding.

Just wanted to confirm, it is indeed should be 2 probes by default, right?

-d

Reply via email to