Orange Forum • View topic - discretization problem on small values or intervalls

discretization problem on small values or intervalls

A place to ask questions about methods in Orange and how they are used and other general support.

discretization problem on small values or intervalls

Postby silvio » Thu Jul 29, 2010 17:00

Hi,

my values of my featureset (it consists of 500 features) are tf-idf weights. So I have many very very little values like 0.000xx for "." or 0.00000.xx for "-".

When I now want to apply AssociationRules it gives me the following error:
Code: Select all
orange.AssociationRulesInducer(data, supp = 0.3, classificationRules = 1)


orange.KernelException: 'orange.AssociationRulesInducer': cannot induce rules with non-discrete attributes (such as '.')

When I call
Code: Select all
orange.EntropyDiscretization(".", data)
I get
'C' as an result. What should 'C' mean?

Thanks for your hints.


Regards,

Silvio

Return to Questions & Support