Is it efficient to use polynomial regression for multiple features like x, y and more…Such that we would have the parameters like x, x2, x3 and y, y2, y3 and more ?
This depends on what you mean by “efficient”.
Typically you’ll only create additional features for the input.
There is no benefit to creating new polynomial features for the output.
We’re trying to create a model that predicts the ‘y’ outputs. There is no reason to create different ‘y’ outputs.
Oh understood, Thank You!