Summary

At a high level, in this chapter you learned about four of the most popular classifiers out there: k-Nearest Neighbors, logistic regression, decision trees, and random forests. Not only did you learn the basics and mechanics of these four algorithms, but you saw how easy they were to perform in R. Along the way, you learned about confusion matrices, hyper-parameter tuning, and maybe even a few new R incantations.

We also visited some more general ideas; for example, you've expanded your understanding of the bias-variance trade-off, looked at how the GLM can perform great feats, and have become acquainted with ensemble learning and bootstrap aggregation. It's also my hope that you've developed some intuition as to which classifiers to use in different situations. Finally, given that we couldn't achieve perfect classification on our diabetes dataset, I hope that you've gained an appreciation for the art and difficulty of classification. Perhaps you've even caught the statistical learning...