Laying Linear Regression Foundations

Summary

Deep learning uses data inputs to guess the nonlinear plane that will most correctly go through the middle of a set of data points in a more sophisticated manner than linear regression. This chapter discusses variables and how to work with them to create a linear regression. Regression boasts a long history in different domains: statistics, economics, psychology, social sciences, and political sciences. Using the gradient descent algorithm, linear regression can find the best set of beta coefficients to minimize a cost function given by the squared difference between the predictions and the real values. The easiest way to model complex relations is by employing mathematical transformations of the predictors using polynomial expansion. A solution to a problem involving a binary response would be to code a response vector as a sequence of ones and zeros. The chapter focuses on solving overfitting using selection and regularization.