Dive Into Data Science

5 Binary Classification

Many difficult questions can be phrased simply as yes/no questions: To buy the stock or not? To take the job or not? To hire the applicant or not? This chapter is about binary classification, the technical term for answering yes/no questions, or deciding between true and false, 1 and 0.

We’ll start by introducing a common business scenario that depends on binary classification. We’ll continue by discussing linear probability models, a simple but powerful binary classification approach based on linear regression. We’ll also cover logistic regression, a more advanced classification method that improves on some of the ...