Accelerated Hierarchical Density Based Clustering

By Leland McInnes, John Healy. 2017 IEEE International Conference on Data Mining Workshops (ICDMW), 2017

519 citations recorded by the source corpus.

We present an accelerated algorithm for hierarchical density based clustering. Our new algorithm improves upon HDBSCAN*, which itself provided a significant qualitative improvement over the popular DBSCAN algorithm. The accelerated HDBSCAN* algorithm provides comparable performance to DBSCAN, while supporting variable density clusters, and eliminating the need for the difficult to tune distance scale parameter epsilon. This makes accelerated HDBSCAN* the default choice for density based clustering.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Signal Processing

View Paper: https://www.semanticscholar.org/paper/ddaa43040c2401bf361accac952497e3a58f5a3b

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Maarten Grootendorst
2. Huili Chen
3. Hossein Yalame
4. Azalia Mirhoseini
5. D. Angelov
6. Hossein Fereidooni
7. Shaza Zeitouni
8. Farinaz Koushanfar

In Computer Vision and Pattern Recognition

1. Huili Chen
2. Hossein Yalame
3. Azalia Mirhoseini
4. D. Angelov
5. Farinaz Koushanfar
6. Maarten Grootendorst

In Signal Processing

1. Huili Chen
2. Hossein Yalame
3. Azalia Mirhoseini
4. D. Angelov
5. Farinaz Koushanfar
6. Maarten Grootendorst

Citations in the corpus

• FLAME: Taming Backdoors in Federated Learning (2022, methodology): “In the following, we describe HDBSCAN in detail, following the implementation of McInnes et al. [36, 37].”
• Top2Vec: Distributed Representations of Topics (2020, methodology): “In order find the dense areas of documents in the semantic space, density based clustering is used on the document vectors, specifically Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN) [25, 26, 27].”
• Exploring spatio-temporal changes of city inbound tourism flow: The case of Shanghai, China (2020, methodology): “(1) HDBSCAN HDBSCAN algorithm is a combination of hierarchical clustering and density clustering, which uses variable mutual reach distance to separate clustering results of different densities from sparse noise points (Campello et al., 2013; Mcinnes & Healy, 2017; McInnes et al., 2017).”
• BERTopic: Neural topic modeling with a class-based TF-IDF procedure (2022, methodology): “For fair comparisons between BERTopic and Top2Vec, the parameters of HDBSCAN and UMAP were fixed between topic models.”