Our paper, “Clustering Validation with The Area Under Precision-Recall Curve,” co-authored with Ivan Gesteira Costa, has been accepted at the 19th International Conference on Similarity Search and Applications (SISAP). The conference will be held in Brno, Czech Republic, from October 14 to 16, 2026.

The study investigates the use of the Area Under the Precision-Recall Curve (AUPRC) and related metrics for relative clustering validation. While the Area Under the ROC Curve for Clustering (AUCC) laid important groundwork, it can be misleading in scenarios characterized by significant cluster size imbalance. Our findings demonstrate that AUPRC is not only appropriate for clustering validation but should be preferred in these unbalanced scenarios, advancing the development of a unified performance evaluation framework for both supervised and unsupervised learning models.

This research was carried out during my period as a visiting scholar at RWTH Aachen University. With thanks to the German Academic Exchange Service (DAAD) for financial support under the Research Stays for University Academics and Scientists 2022 Program, Grant #91835520.

As soon as published, the paper will be linked at the Publications page.