[CONCEPTUAL FLAW] What this package calls "DBCVI" has nothing to do with DBCVI
The density-based clustering validity index is described here:
Moulavi, Davoud, et al. "Density-based clustering validation." Proceedings of the 2014 SIAM international conference on data mining. Society for Industrial and Applied Mathematics, 2014.
I've attached a copy of the PDF for your reference.
I'm not sure where the metric you're computing with the "DBCVI" method originates from, but it has nothing to do with the actual DBCV index, which is specifically designed to avoid reliance on centroids, is density rather than distance based, and anges on [-1,1] with higher scores being better.
[Moulavi et al. - 2014 - Density-Based Clustering Validation.pdf](https://github.com/user-attachments/files/23318402/Moulavi.et.al.-.2014.-.Density-Based.Clustering.Validation.pdf)
0 条评论