Member-only story
What is a Confusion Matrix?
What are they, why, and when should we use them?
Introduction
In machine learning, evaluating the performance of a model is just as important as developing the model itself. The most popular approach data scientists implement when evaluating a model's performance is a confusion matrix.
But what exactly is a confusion matrix, and why is it so widely used? In this article, we’ll break down the concept, explain its components, and show you when and how to use it.
What Is a Confusion Matrix?
A confusion matrix is a model performance measurement tool for machine learning classification problems. Its output compares model predictions against true labels, creating a visual representation of the number of correct and incorrect predictions, categorized by each class in the dataset.
This matrix is especially beneficial when tackling binary classification problems (e.g. “spam emails” vs. “not spam emails”), but can also be implemented when tackling multi-class problems.
Confusion Matrix Structure
The confusion matrix is typically a 2x2 table for binary classification, structured like the following:
Each cell in the table has a specific meaning:
- True Positive (TP): The model correctly predicted the positive class.
- True Negative (TN): The model correctly predicted the negative class.
- False Positive (FP): The model incorrectly predicted the positive class (a “false alarm”). This is also known as a Type I Error — rejecting a true null hypothesis.
- False Negative (FN): The model incorrectly predicted the negative class (a “miss”). This is also known as a Type II Error — failing to reject a false null hypothesis.
Metrics Derived from the Confusion Matrix
Using a confusion matrix you can calculate a variety of evaluation metrics to understand your model’s performance better…
