In this chapter, we talked about how to improve various classifiers by combining them into an ensemble. We discussed how to average the predictions of different classifiers using bagging and how to have different classifiers correct each other's mistakes using boosting. A lot of time was spent discussing all possible ways to combine decision trees, be it decision stumps (AdaBoost), random forests, or extremely randomized trees. Finally, we learned how to combine even different types of classifiers in an ensemble by building a voting classifier.

In the next chapter, we will talk more about how to compare the results of different classifiers by diving into the world of model selection and hyperparameter tuning.