DelTriC: A Novel Clustering Method with Accurate Outlier

TL;DR

DelTriC is a new clustering algorithm that combines low-dimensional projection, Delaunay triangulation, and back-projection to improve clustering accuracy and outlier detection in high-dimensional data.

Contribution

It introduces a novel clustering approach that decouples neighborhood construction from clustering decisions using triangulation and back-projection, enhancing robustness and scalability.

Findings

Outperforms traditional clustering methods like k-means, DBSCAN, HDBSCAN.

Scalable and accurate in high-dimensional spaces.

Significantly improves outlier detection capabilities.

Abstract

The paper introduces DelTriC (Delaunay Triangulation Clustering), a clustering algorithm which integrates PCA/UMAP-based projection, Delaunay triangulation, and a novel back-projection mechanism to form clusters in the original high-dimensional space. DelTriC decouples neighborhood construction from decision-making by first triangulating in a low-dimensional proxy to index local adjacency, and then back-projecting to the original space to perform robust edge pruning, merging, and anomaly detection. DelTriC can outperform traditional methods such as k-means, DBSCAN, and HDBSCAN in many scenarios; it is both scalable and accurate, and it also significantly improves outlier detection.