Soccer Analytics with Machine Learning

Chapter 6. Advanced Classification Methods

A Note for Early Release Readers

With Early Release ebooks, you get books in their earliest form—the author’s raw and unedited content as they write—so you can take advantage of these technologies long before the official release of these titles.

This will be the 6th chapter of the final book. Please note that the GitHub repo will be made active later on.

If you’d like to be actively involved in reviewing and commenting on this draft, please reach out to the editor at [e-mail].

In Chapter 5, we built the core classification workflow using logistic regression and KNN. Those models are valuable starting points, but they also impose constraints: logistic regression is linear in its decision boundary, and KNN can become unstable or hard to scale as feature spaces grow.

This chapter introduces tree-based models as a more flexible extension of that work. Decision trees, random forests, and gradient boosting methods can represent nonlinear relationships, interactions, and threshold effects that arise naturally in soccer data.

That flexibility is especially useful in soccer analytics, where performance often depends on combinations of context: distance and angle together, pressure and location together, possession state and game situation together. Tree-based models can capture those structures without requiring us to write every interaction by hand.

At the same time, tree-based models preserve an important link to interpretation. ...