6. Regression Analysis

In this chapter, we introduce regression analysis and some of its applications in data science using Python tools. We show how regression analysis allows us to understand the behavior of data better, to predict data values (continuous or discrete), and to find important variables by means of building a model from the data. We present four different regression models: simple linear regression, multiple linear regression, polynomial regression and logistic regression. We also emphasize the properties of sparse models in the selection of variables. We use different Python toolboxes to build and apply regression models with ease. Specific visualization tools from Seaborn allow qualitative evaluation; while tools from the Scikit-learn library make quantitative evaluation easier, computing several validation measures. Depending on our aim, visual inspection of the data, statistical analysis or prediction, we chose one tool or another. Regression models are motivated by three real problems that deal with the following questions. Is the climate really changing? Can we predict the price of a new market, given any of its attributes? How many goals makes a football team the winner or the loser?