Orange Forum • View topic - Neural Network Error

Neural Network Error

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

Neural Network Error

Postby JoshuaR » Fri Apr 04, 2014 18:43

I have data with discrete 1/0 values for some twenty-ish attributes, and then a class attribute that is also 1/0 that I'm trying to predict/learn. There aren't any missing data values. The table is full of just 1s and 0s. I get this error when I connect a neural network to the Test Learners widget:

Code: Select all
ValueError                                    Traceback (most recent call last):
  File "C:\Users\jr211\Documents\Anaconda\lib\site-packages\Orange\OrangeCanvas\scheme\widgetsscheme.py", line 603, in process_signals_for_widget
    widget.handleNewSignals()
  File "C:\Users\jr211\Documents\Anaconda\lib\site-packages\Orange\OrangeWidgets\Evaluate\OWTestLearners.py", line 583, in handleNewSignals
    self.score(filter(needsupdate, self.learners))
  File "C:\Users\jr211\Documents\Anaconda\lib\site-packages\Orange\OrangeWidgets\Evaluate\OWTestLearners.py", line 352, in score
    if (multilabel and isinstance(learner, Orange.multilabel.MultiLabelLearner)) or predictor(new[0]).varType == new.domain.classVar.varType:
  File "C:\Users\jr211\Documents\Anaconda\lib\site-packages\Orange\classification\neural.py", line 304, in __call__
    mt_value = Orange.data.Value(self.domain.class_var, cprob.values().index(max(cprob)))
ValueError: nan is not in list


Any ideas?

Return to Questions & Support