Orange Forum • View topic - Widget predictiond doesn't predict

Widget predictiond doesn't predict

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

Widget predictiond doesn't predict

Postby sapienza » Mon Oct 01, 2007 23:53

Hi i have used the widget of CN2 to produce rules for a dataset with discrete and categorical values, and to this dataset i have applied a data sampler. Now i wanna use the widget predictions to use the rules produced with the sampled data (test set) to predict the behaviour of the train data, which are the remaining examples output from the data sampler. So i have done these links:
- dataset to data sampler
- data sampler (examples) to CN2
- data sampler (remaining examples) to predictions
- CN2 (classifier) to predictions

Unfortunatel this doesn't work, and in the widget predictions nothing appears. Is this a bug or is it just me using that widget in a wrong way?If it's me can someone please explain me how do i have to link the widgets to have a prediction?

Thanks everybody

Postby sapienza » Tue Oct 23, 2007 10:28

It happened again. Here is the error:

FIle: OWBaseWiget.py in line 529
Function name:processSignals
Code: handler (value, (widgetFrom, name From, id))
File: OWPredictions.py in line 317
Function name: setClassifier
Code: if self.outvar.varType == orange.VarTypes.Continuous:
Exception type:
Exception value: 'NoneType' object has no attribute 'varType'


Return to Questions & Support