Can someone help me how space transformation works on linear regression problems because I have been confused.
When we perform space transformation with a function e.g. $\varphi (x)$ we perform the transformation on dependent variable, independent or both? with which method?
For example if I have the following nonlinear model $y = 2x_1^2 + 8x_2^2 + 4x_1x_2 + 8x_2 + 9x_1$ and I want to transform to a space where the problem of estimating the model is linear?
Also any link related to the problem.
Thanks