hdbscan: Hierarchical density based clustering

By Leland McInnes, John Healy, S. Astels. Journal of Open Source Software, 2017

2,775 citations recorded by the source corpus.

HDBSCAN: Hierarchical Density-Based Spatial Clustering of Applications with Noise

Subfields: Artificial Intelligence, Epidemiology

View Paper: https://www.semanticscholar.org/paper/d4168c0480bc8e060599fe954de9be1007529c93

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

In Epidemiology

Citations in the corpus

Untangling the Galaxy. I. Local Structure and Star Formation History of the Milky Way (2019, methodology): “We perform a clustering analysis on this dataset using Python implementation of HDBSCAN (Hierarchical Density-Based Spatial Clustering of Applications with Noise, McInnes et al. 2017).”
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].”
Interpreting Word-Level Hidden State Behaviour of Character-Level LSTM Language Models (2018, positive): “We used the Python implementation of McInnes et al. (2017).”
Statistical power for cluster analysis (2020, methodology): “A class of algorithms that does not require the prespecification of an expected number of clusters includes DBSCAN [45] and HDBSCAN [34].”
Robust Preference Learning for Storytelling via Contrastive Reinforcement Learning (2022, methodology): “Since HDBSCAN works best on vectors of low dimensionality projection to a low dimension is an important preprocessing step.”
Accelerated Hierarchical Density Based Clustering (2017, methodology): “Furthermore, in both the 2-dimensional and 50-dimensional cases, the accelerated Python version demonstrates roughly two orders of magnitude better absolute run-time performance on data set sizes of 200,000 points.”
Security and Privacy Challenges of Large Language Models: A Survey (2024, neutral): “By leveraging the capabilities of HDBSCAN [158], CUBE aims to provide an effective means of differentiating clusters associated with both normal and poisoned data.”
Finding, visualizing, and quantifying latent structure across diverse animal vocal repertoires (2020, methodology): “…projections of spectrograms (HDBSCAN is applied to 100-dimensional PCA projections rather than spectrograms directly because HDBSCAN does not perform well in high-dimensional spaces [66]), k-means [67] clustering applied over UMAP, and k-means clustering applied over spectrograms (Fig 10; Table 1).”
Unspeech: Unsupervised Speech Context Embeddings (2018, methodology): “We use HDBSCAN [33], a modern hierarchical density-based clustering algorithm for our experiments since it scales very well to a large number of utterances and the number of clusters does not need to be known apriori.”
Fast and flexible bacterial genomic epidemiology with PopPUNK (2018, methodology): “The alternative approach uses HDBSCAN to classify a subsample of 10 5 points using the Boruvka ball tree algorithm (McInnes et al. 2017) .”
Evaluating Large Language Models in Generating Synthetic HCI Research Data: a Case Study (2023, neutral): “Intuitively, measur w (4) Cluster precision es ho large the dimensionality-reduced embedding vectors using portion of the generated of data samples lie close eal HDBSCAN the popular DBSCAN algorithm r [49], a variant to data, and recall measures how large portion of the real data 22 co [ is vered by”
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].”
Evaluating Merging Strategies for Sampling-based Uncertainty Techniques in Object Detection (2018, background): “sion of DBSCAN, a density-based clustering algorithm, that is robust to parameter selection and allows for clusters of varying densities [29].”
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.”
A Large-Language-Model Framework for Automated Humanitarian Situation Reporting (2025, unknown): “Starting from the collected documents, our pipeline proceeds in six main stages: • Semantic text clustering: the documents are first segmented into smaller text units and grouped into clusters using HDBSCAN 20 .”