This is an assignment question. Can someone someone give me some clue on how to get through:

A generative approach is used for a binary classification problem and it turns out that the resulting classifier predicts positive at all points x in the input space. What can we conclude for sure? Check all that apply.

  • There are no negative points in the training set.
  • The positive points are spread out over the space, while the negative points are concentratd in a small region.
  • There are fewer negative points than positive points in the training set.
  • The density of positive points is greater than the density of negative points everywhere in the space.

You need to select points that apply out of those four bullet points. Thanks

  • $\begingroup$ Are we missing some more information? $\endgroup$ – Aditya Mar 4 '18 at 7:07
  • $\begingroup$ 'There are fewer points than points in the training set' ... are you sure you got that right? $\endgroup$ – Elias Strehle Mar 4 '18 at 9:18
  • $\begingroup$ There are some edits to the questions, Thanks - @Aditya $\endgroup$ – Developer Mar 4 '18 at 21:20
  • $\begingroup$ Going wiith option B,C? Even A can be True( worst case)? $\endgroup$ – Aditya Mar 5 '18 at 1:41
  • $\begingroup$ I have tried a lot of combinations but nothing work (got wrong each time) - @Aditya $\endgroup$ – Developer Mar 5 '18 at 2:08

While training your binary classifier, check for class imbalance. That's the only way you can take any final decision about the test data. As you don't know the actual result of test data, therefore, you need to make sure the classifier is good enough to generalize any new data.

|improve this answer|||||

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.