AIExplainer
Machine Learning Intermediate 2 min read

What is area under the PR curve?

A measure of a model's performance in classification tasks

The area under the PR curve, also known as AUPRC, is a metric used to evaluate the performance of a classification model. It plots the precision of the model against its recall, and the area under this curve represents the model's ability to distinguish between positive and negative classes.

Imagine a treasure hunter who wants to find all the treasure chests in a large field. The area under the PR curve is like a measure of how well the treasure hunter can find all the chests without digging up too much dirt. The precision is like the ratio of chests to dirt dug up, and the recall is like the ratio of chests found to total chests. The area under the curve shows how well the treasure hunter balances these two ratios.

In medical diagnosis, the area under the PR curve can be used to evaluate the performance of a model that predicts the presence or absence of a disease. For example, a model that predicts the presence of cancer based on patient symptoms and test results can be evaluated using the area under the PR curve to determine its accuracy.

The area under the PR curve is used to evaluate the performance of classification models, especially in cases where the classes are imbalanced. It provides a single metric that summarizes the model's performance across different thresholds.

One common misconception is that the area under the PR curve is the same as the area under the ROC curve. While both curves are used to evaluate model performance, they have different x-axes: the PR curve plots precision against recall, while the ROC curve plots true positive rate against false positive rate.

The concept of the area under the PR curve has been around since the 1960s, but it gained popularity in the 2000s with the rise of machine learning and data mining.

AUPRC area under the precision-recall curve precision-recall AUC

Three products for different needs — explore what’s relevant to you.