I'm reading about Skip-Gram and how it represents words as vectors. If I'm correct it not just represents a single words as a vector, but the word and it's neighbours.

So if I have the dataset "the quick brown fox jumped over the lazy dog", Skip-Gram (with a window of 1) learns vectors for ([the, brown], quick) (where the words between [] represent the neighbours), ([quick, fox], brown), ... rather than for the single words "quick" and "brown".

Is this correct, and if not how does it use the window/neighbours?

  • $\begingroup$ en.wikipedia.org/wiki/N-gram#Skip-gram $\endgroup$ – D.W. Apr 4 '16 at 2:22
  • $\begingroup$ I read that, but didn't fully understand. So that's why I'm asking here for clarification. $\endgroup$ – The Oddler Apr 8 '16 at 11:18
  • $\begingroup$ OK, so you didn't fully understand. What specifically about Wikipedia was confusing or unclear? Wikipedia gives you an example. Why not work through that example and explain exactly why you are confused? That would make it easier to help you and addres any misconceptions. You are proposing what the answer might be for your dataset. Where did you get that from, and why do you think it's correct? What you write looks nothing like Wikipedia's example, so it's not clear where you're getting that from. $\endgroup$ – D.W. Apr 8 '16 at 15:26

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.