The threshold is interpreted with respect to the
attribute evaluator being used. For example, if you
use the InfoGainAttributeEval then a threshold that
you set is the minimum information gain (between
an attribute and the class) that is acceptable. All
attributes with scores less than the threshold are
[mailto:email@example.com]On Behalf Of David A.
Sent: Wednesday, 27 February 2002 6:48 a.m.
Subject: [Wekalist] Ranker
When using weka.attributeSelection.Ranker, what does the threshold really
mean? That is, how does it relate to my variables? Is it tested against the
actual value or against some other measure.
Wekalist mailing list