A Density-Based Algorithm for Discovering Clusters in Large Spatial Databases with Noise

By Martin Ester, Hans-Peter Kriegel, Jörg Sander, Xiaowei Xu. 1996

How researchers see this paper

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

In Model Architectures

Citations in the corpus

Local Aggregation for Unsupervised Learning of Visual Embeddings (2019, methodology): “Given the need for cluster-based identiﬁcation, a variety of alternative approaches to k -means are theoretically possible, including DBSCAN [16], Afﬁnity Propagation [18], spectral methods [59, 62], and gaussian mixtures [56, 57].”
Unsupervised and Semi-supervised Learning with Categorical Generative Adversarial Networks (2015, background): “It should be mentioned that there exist clustering algorithms – such as DBSCAN (Ester et al., 1996) or spectral clustering methods – which can correctly identify the clusters in the datasets by making additional assumptions on the data distribution. k - m e a n s R I M + NN data + class assignment”
The Dangers of Post-hoc Interpretability: Unjustified Counterfactual Explanations (2019, methodology): “As mentioned earlier, an easy way to implement this is to use the clustering algorithm DBSCAN on B f ( a 0 ) x ∪{ a 0 } with parameter values (cid:15) and minPts = 2 .”
Representation Learning on Graphs: Methods and Applications (2017, background): “Again, because each node is associated with real-valued vector embedding, it is possible to apply any generic clustering algorithm to the set of learned node embeddings ( e.g. , k-means or DB-scan [22]).”
Predicting multiple conformations via sequence clustering and AlphaFold2 (2023, positive): “To do this, we made predictions for DBSCAN-generated KaiB clusters in the model MSA Transformer 36 using its default settings.”
To Trust Or Not To Trust A Classifier (2018, neutral): “Level-set estimation has a long history [37, 38, 39, 40, 41, 42].”
End-to-End Text Recognition with Hybrid HMM Maxout Models (2013, methodology): “Since MSERs would ideally correspond to character regions, we form candidate line boxes by clustering the character candidates with DB-SCAN (Ester et al., 1996) using multiple distances to obtain candidate line-level bounding boxes.”
CUDA-Based Parallelization of Power Iteration Clustering for Large Datasets (2017, methodology): “[13] presented the algorithm G-DBSCAN, a GPU parallel version of the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) [14].”
A Survey of Collaborative Filtering Techniques (2009, unknown): “DBSCAN [79] and OPTICS [80] are well-known density-based clustering methods.”
Mining, indexing, and querying historical spatiotemporal data (2004, methodology): “DB-SCAN [2] has quadratic cost to the number of clustered points, unless an index (e.g., R–tree) is available.”
Automatic Subspace Clustering of High Dimensional Data (2005, methodology): “Recent work in the database community includes CLARANS (Ng and Han, 1994), Focused CLARANS (Ester et al., 1995), BIRCH (Zhang et al., 1996), DBSCAN (Ester et al., 1996) and CURE (Guha et al., 1998).”
Ensemble learning for credit card fraud detection (2018, result): “Rather, the they are more or less uniformly distributed in all the clusters. we also attempted other unsupervised clustering techniques such as DBSCAN [8], and a mixture of gaus-sians [19], but obtained similar results.”
Efficient Nearest Neighbor Language Models (2021, background): “neighbors unlike typical density-based clustering methods (Ester et al., 1996) which may amplify errors.”
Traffic classification using clustering algorithms (2006, methodology): “In this paper, we confirm their observations by evaluating two clustering algorithms, namely K-Means [7] and DBSCAN [5], that to the best of our knowledge have not been previously applied to this problem.”
Interpretable dimensionality reduction of single cell transcriptome data with deep generative models (2017, methodology): “By clustering the two-dimensional t-SNE embedding using DBSCAN 61 — a density-based clustering algorithm, the authors identi ﬁ ed 39 clusters after merging the clusters without enough differentially expressed genes between any two clusters.”