Open Conference Systems, CLADAG2023

Font Size: 
Empirical Analysis of the Quadratic Scoring for Selecting Clustering Solutions
Luca Coraggio, Pietro Coretto

Last modified: 2023-06-22

Abstract


Selecting an optimal clustering solutions is a difficult problem, and there exist many data-driven validation strategies to perform this task. In this paper, we focus on a recent proposal, the BQH and BQS criteria, based on quadratic discriminant scores and bootstrap resampling. We provide more insight on these criteria, comparing them with a likelihood-based alternative and using different resampling schemes.