Summary

In the context of machine learning, we train a model and test it to predict an outcome. In this chapter, we had an in-depth look at the simple yet extremely effective methods of linear regression and MARS to predict a quantitative response. We also applied the data preparation paradigm put forth in Chapter 1, Preparing and Understanding Data, to quickly and efficiently get the data ready for modeling. We produced several simple plots to understand the response we were trying to predict, explore model assumptions, and model results.

Later chapters will cover more advanced techniques like Logistic regression, Support Vector Machines, Classification, Neural Networks, and Deep Learning but many of them are mere extensions of what we've learned in this chapter.