0
$\begingroup$

I was asked in an interview question, can a perceptron classifier ever reach 100% accuracy on some kind of non linearly seprabale training data in 2D. I said that no it can't because the data is not linearly seprabale. We then proceed to the next question, but maybe its just me, but from the look of the interviewr, I'm not sure that was the right answer. Am I missing something or just paranoid? Can the perceptron reach 100% accuracy on any training data?

Thanks!

$\endgroup$
0
$\begingroup$

A single-layer perceptron can only learn linearly separable patterns (with 100% accuracy), as noted in the Wikipedia article on perceptrons. (Advice for the future: check standard references before asking here.)

A multilayer feedforward neural network (sometimes called multilayer perceptron) can learn other patterns.

If someone just says "perceptron", in my eyes it is at least a little ambiguous whether that means a single-layer perceptron or possibly could encompass multi-layer networks.

$\endgroup$
3
  • $\begingroup$ Yeah I know that but what if we project the data to a higher dimension for example? $\endgroup$
    – mathkid
    Nov 6 '20 at 23:31
  • $\begingroup$ @mathkid, that wasn't in your question. Please make sure that your post includes everything you want answered. This site doesn't work very well for interactive conversations or back-and-forth. Perhaps it is worth asking a new question where you ask about that scenario? $\endgroup$
    – D.W.
    Nov 7 '20 at 2:20
  • $\begingroup$ Yeah sorry for this. I thought that I explained it on my post but I guess not. $\endgroup$
    – mathkid
    Nov 7 '20 at 2:22

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.