Machine Learning, Illustrated: Evaluation Metrics for Classification

A comprehensive (and colorful) guide to everything you need to know about evaluating classification models

I realized through my learning journey that I’m an incredibly visual learner and I appreciate the use of color and fun illustrations to learn new concepts, especially scientific ones that are typically explained like this:

From my previous articles, through tons of lovely comments and messages (thank you for all the support!), I found that several people resonated with this sentiment. So I decided to start a new series where I’m going to attempt to illustrate machine learning and computer science concepts to hopefully make learning them fun. So, buckle up and enjoy the ride!

Let’s begin this series by exploring a fundamental question in machine learning: how do we evaluate the performance of classification models?

In previous articles such as Decision Tree Classification and Logistic Regression, we discussed how to build classification models. However, it’s crucial to quantify how well these models are performing, which begs the question: what metrics should we use to do so?

To illustrate this concept, let’s build a loan repayment classification model.