Becoming a Data Head

CHAPTER 9Understand the Regression Model

“Regression analysis is like one of those fancy power tools. It is relatively easy to use, but hard to use well—and potentially dangerous when used improperly.”

Charles Wheelan in Naked Statistics¹

SUPERVISED LEARNING

The preceding chapter showcased unsupervised learning—discovering patterns or clusters in a dataset without predefined groups. Remember, in unsupervised learning, we come with no preconceived notions or patterns. We instead take advantage of the underlying aspects of data, establish some guardrails, and then let the data organize itself.

However, there will be many occasions when you do in fact know something about the underlying data. In that case, you'll need to use supervised learning to find relationships in data with inputs and known outputs. Here, you have the correct answers to “learn” from. You can then judge the robustness of the model against what you know about the real world. A good model will allow you to make accurate predictions and provide some explanation to the underlying relationships between the data's inputs and outputs.

Supervised learning, as you might recall, made an early cameo in the book's Introduction, back when you first started your Data Head journey. We asked you to predict if a new restaurant location would be a chain or independent restaurant. To make your guess, you first learned from existing restaurant locations (inputs) and known labels of “chain” or “independent” (outputs). You ...