I am using the software Orange to undertake a random forest classification of geo-chemical data. I am trying to classify points as 0 or 1 based on whether it is a mineral occurrence or not.

My training data returns decent results based on test scores and a confusion matrix, but my test dataset returns 100%. I have a feeling it's because some of training data is being used as test data, but I don't know how to fix it. Does anyone here use Orange. Can you take a look at my workflow? I can't share any more due to confidentiality.

enter image description here

  • $\begingroup$ What would be the point of test data, if you already used it to train the network? $\endgroup$ – Stephen Rauch May 24 at 1:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.