I recently finished an ML assignment in which I repeatedly tested various models to determine which one would work best for my data.
To select the best model, I looked at metrics like MSE, MAE, R^2, etc. And I went with the option that, according to these measures, produced the best outcome.
Is this also the way businesses operate? or are there other elements that must be taken into account?
I assume that persuading stakeholders to adopt this model, the technology needed to train and retrain the model, and the volume of data required to provide decent results for a specific model will be some of the issues, but are there any others? I just want to know how businesses go about this process.