0
$\begingroup$

Suppose a machine learning model is trained on two features (f1, f2). If there comes a new feature(f3) is there any way to pass this new feature to existing model and retrain instead of training a model again on the three features?

$\endgroup$
2
$\begingroup$

So the question asks if you can retrain the model after adding another feature.

Answer is yes but you need to instantiate a new model which can accept a 3-feature/dimensional input.

This is due to the mechanics of neural networks whereby in the 1st layer, the features in your original model would be multiplied by a ($2 x n$) matrix where $n$ represents the number of hidden neurons in the second layer of the model. Therefore by multiplying a ($3 x 1$) input with the weight matrix will throw an error.

$\endgroup$

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.