Before we get down to brass tacks on how we will tackle building regression models using TDD, we need to refresh ourselves on some of the finer points. Multiple regression comes packed with some assumptions and different measures of model quality. A good amount of this information can be found in A Second Course in Statistics Regression Analysis, Mendenhall & Sincich, Pearson (2011).
When a lot of people are introduced to regression, their main take-away is this is how we draw a line through our data to predict what it will be. To be fair, that's pretty accurate, but there's a fair amount of nuance in this that we need to explicitly discuss.
First let's discuss the standard multiple regression model form. It looks like this:
Here y is our dependent variable. Every x variable is an independent variable. y being a dependent variable means it is dependent on the values of the independent variables and the error term . The error term is...