Interview Query

Separate Models for Age Groups

Have you seen this question before?

Suppose you work for a media company.

You are training a Boosting algorithm on the user activity dataset to predict whether users will convert to a paid subscription.

When training your model, a co-worker suggests that the model should be separated into two models, one for older users and another for new users.

Is this a good approach? Why or why not?

Next question: Deer Density
Loading comments