1
$\begingroup$

I want perform demand forecast for particular item based on attributes.Did I need to train the model with unsold items ? by maintaining sales Quantity as zero or go with only items sold in training period.

$\endgroup$
1
$\begingroup$

In short, if there is supply, it should be in the model to determine demand:

  • If you have XXXL pink jeans in your store and don't sell them, I think it needs to be in the model.
  • If you don't have the pink jeans in your store for sale, it doesn't.

It gets interesting when you'd sell some XS pink jeans and blue XXXL ones, but you'd at least have to offer them in the store to be able to sell them and get measurement data to train the model.

| improve this answer | |
$\endgroup$
  • $\begingroup$ This is not exactly true. Even if you don´t offer them, as long as you get a demand measurement, you could do some king of prediction. $\endgroup$ – Neoromanzer Nov 10 '17 at 20:22
1
$\begingroup$

If you want to estimate demand, no, you do not need to train the model with unsold items for which you have supply, since they do not affect it. Having an excess supply/capacity does not affect your customer behaviour.

If you have supply, but no demand, it needs to be there.

However, also take into account that if you have had a demand greater than the supply at any moment, you have to do proper demand unconstraining to have a good forecast of your demand.

| improve this answer | |
$\endgroup$

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.