Introduction

In the preceding chapter, we saw how to solve data problems using unsupervised learning algorithms and applied the concepts that we learned about to a real-life dataset. We also learned how to compare the performance of various algorithms and studied two different metrics for performance evaluation.

In this chapter, we will explore the main steps for working on a supervised machine learning problem. First, this chapter explains the different sets in which data needs to be split for training, validating, and testing your model. Next, the most common evaluation metrics will be explained. It is important to highlight that, among all the metrics available, only one should be selected as the evaluation metric of the study, and its selection should be made by considering the purpose of the study. Finally, we will learn how to perform error analysis, with the purpose of understanding what measures to take to improve the results of a model.

The previous concepts apply to...