Let's say, we have a model M1 that is trained on a particular device D1 (let's say the device is a Car), with historical data over 6 months of D1. M1's purpose is to predict the behavior of the car. Now we have another device D2 (another car), with same specification but new and yet to be run.

Can we use M1 to predict the behavior of D2, since we do not have any data on D2?


1 Answer 1


The answer to your question is yes, essentially while building a model what we are doing in reality is devising a mathematical function of a physical object (here car). So essentially what we are building is a mathematical model of a car.

What you should be aware of is ? 1. Say for e.g. if you are building a model for a car A and Car B, and if car A is a gasoline(internal combustion car) and car B is an electric car, then things become tricky if you want to build a mathematical function of internals of a car. 2. This does not affect if the model does not consider the internals of the car meaning the outcome (prediction) is not dependent on the internals of the car, which may not be the case always.

but in general you can use the model, it is always preferable to do a validation on the new dataset.


Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.