Which loss function is appropriate in my model?

First of all, my model is a discrete model, and it is not a classification model.

I’ve seen a lot of loss functions, but most of them are classification models.

Besides, in addition to the classification model, the MSE loss function is generally used.

① Did anyone recommend the loss function to me?
② Can RNN be used in a discrete model?

Example:There are some data about BMI(body mass index)
(Although It’s not my model, but The characteristics are very similar.

BMI = kg/(m^2)

Data Sample:

If I understand you correctly, your model must take cm and kg inputs and calculate the target, right?

Why do you say the model is discrete? The targets are floats, so MSE loss should be fine.
The cm and kg inputs are integers, but converting them to floats shouldn’t pose any problems.

I don’t see how an RNN could be used in this situation.

Thank you for your reply.

Yes, you understand correctly, but the mean of not discretee model is that It is not time series model like stock price prediction.