In this chapter, we introduced important models for classification tasks and we used them in practice with the credit card default dataset. We covered the most commonly used models in the application and research industry, looking at three types of classification tasks—binary, multiclass, and multilabel classification. We learned about the logistic regression model, which tries to estimate the conditional probability of an observation belonging to the positive class. Toward the end of the chapter, we learned how multiclass classification is done automatically by scikit-learn models using the One-versus-All method.
Now that we have learned the basic models for regression and classification tasks, it is time for us to take a look at a family of models that have become very popular in the last years, not only for doing predictive analytics but for their success in...