Let's say I have 1000 participants within a tournament, each of the participants having a single ID. One instance of this tournament is a 1 vs 1 contest when one wins and the other loses.

Now let's say ID11 is facing ID200, I'd like to represent the history and outcomes of the previous contests between ID11 and all of the other participants if they faced each other (sparse). Same for ID200.

example: ID11 won/lost vs ID500, ID600, ID250, etc... ID200 won/lost vs ID600, ID400, ID800, etc... in a single vector. The idea is to be able to use the information of who beat who and therefore each ID should be their own entity to be processed.

I have thought about one-hot encoding this vector but it would become a 2000-length input just for this...

Would anyone have any insights or input or concepts that i could look into to implement such representations?

Thank you very much.

  • $\begingroup$ what is the final goal of such a vector? $\endgroup$
    – noe
    Nov 26, 2022 at 10:49
  • $\begingroup$ See if it can predicts the winner based on those relationships between IDs using a neural network. $\endgroup$
    – Stak33
    Nov 26, 2022 at 10:52


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.