1
$\begingroup$

I used logistic regression to classify 3 classes of images: cars, dogs, cows. The accuracy is 100%, perfectly separated in clusters (scatter plots). When I try to use the learner to predict 3 images (1 dog, 1 car, 1 cow), the predictor simply assign the probability of 1/3 to each image. How can I fix it?

orange3 flows

$\endgroup$
  • $\begingroup$ Is there any way to attach your ows file to your question? Also, the input data set would be helpful. $\endgroup$ – VividD May 26 '17 at 16:18
  • $\begingroup$ this link contains ows and input data jmp.sh/i5dJCZB $\endgroup$ – Max NewSite May 27 '17 at 13:33
1
$\begingroup$

It means that the model you are using fails at classifying the images. Logistic regression may not be the good one for your classification. Try using another model.

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.