Debugging Machine Learning Models with Python

15

Correlation versus Causality

In previous chapters of this book, you learned how to train, evaluate, and build high-performance and low-bias machine learning models. However, the algorithms and example methods we used to practice the concepts that were introduced in this book do not necessarily provide you with a causal relationship between features and output variables in a supervised learning setting. In this chapter, we will discuss how causal inference and modeling could help you increase the reliability of your models in production.

In this chapter, we will cover the following topics:

• Correlation as part of machine learning models
• Causal modeling to reduce risks and improve performance
• Assessing causation in machine learning models