What is Regression?

Regression are fundamental supervised learning methods. It has two categories: linear regression and polynomial regression. The central goal of linear regression is to fit a linear surface to some data. The term ‘linear surface’ refers to a line in two dimensions, a plane in three, and a hyperplane in any more. Polynomial regression is similar, except that the surface it fits to data is not necessarily linear. Both methods are parametric, and training involves minimizing an error function over a set of internal parameters.

Before heading into these methods, you should review the concept of supervised learning. Then we’ll move into theory and usage of linear regression. Finally, we will move to the model choices and cost function of the linear regression. Then we will finish off with the overview of polynomial regression.