0
$\begingroup$

There is Company X

Company X wants to predict if its opportunities will Win or Lose

I trained my model based on the data of company X and did the prediction of new opportunities

Now Company Y and Company Z wants to buy this product

  1. Should I predict their opportunities based on my trained model of (Company X) ?

OR

  1. Do a new model that is trained based on (Company: X+Y+Z)

OR

  1. Do 3 separate models for each company

What is the best approach if I want other companies to use my model? In this case its 3 only, what if I have 20 companies

Thanks

$\endgroup$
0
$\begingroup$

We can't tell you.

If the companies are very similar, then you may be able to transfer a pre-trained model between them. Or take a pre-trained model from X and update it with data from Y, rather than starting from scratch with data from Y.

If the companies are very dissimilar, you may be better off starting from scratch.

Either way, we don't know, and without knowing much more, we can't even make an educated guess. So I recommend you just try all reasonable approaches and see which one outperforms the others on a holdout sample.

Also: make sure everyone involved is fine with this. Company X may not be happy if a model that was trained on their data is sold to someone else. First, they may not want to help a competitor. Second, it is sometimes possible to reverse engineer training data out of a trained model, so there may be issues with confidentiality of training data.

$\endgroup$

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.