Density-Based Clustering over an Evolving Data Stream with Noise

By Feng Cao, Martin Ester, Weining Qian, Aoying Zhou. 2006

1,104 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Signal Processing

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 Signal Processing

Citations in the corpus

Memory Efficient Experience Replay for Streaming Learning (2018, background): “Density-based algorithms cluster data that lie in highdensity regions of feature space together, while labeling points in low-density regions as outliers [8], [10], [15], [29], [47].”
Online Machine Learning in Big Data Streams (2018, methodology): “nt methods are data stream versions of k-means or its variants such as k-medians [237, 35, 80, 177, 104, 9, 239, 96, 143, 3]. Another set of results describes the data stream implementation of DBSCAN [41, 55, 143]. Finally, an online hierarchical clustering algorithm that maintains similarity measur”
Large-scale audio event discovery in one million YouTube videos (2017, methodology): “DenStream operates over a sequence of input vectors x1, . . . , xT , where each xt ∈ Rd.”
Focus: Querying Large Video Datasets with Low Latency and Low Cost (2018, methodology): “Given these requirements, we use the following simple approach for incremental clustering, which has been well-studied in the literature [30, 65].”
Learning to Classify With Incremental New Class (2021, neutral): “However, related problems may cooccur in real-world data streams, e.g., incremental learning [2], concept drift detection [63], and noise data [64].”
The Paradigm-Shift of Social Spambots: Evidence, Theories, and Tools for the Arms Race (2017, methodology): “The approach in [31] considers vectors made of 126 features extracted from both accounts and tweets as input of modiﬁed versions of the DenStream [8] and StreamKM++ [1] clustering algorithms, to cluster feature vectors of a set of unlabeled accounts.”
dbscan: Fast Density-Based Clustering with R (2019, methodology): “Conversely, the label density-based clustering is used for methods which do not assume parametric distributions, are capable of ﬁnding arbitrarily-shaped clusters, handle varying amounts of noise, and require no prior knowledge regarding how to set the number of clusters k .”
A Comprehensive Survey of Clustering Algorithms (2015, methodology): “DenStream, which takes the core idea of the clustering algorithm based on density, is suitable for the nonconvex data set and can deal with outliers efficiently, compared with the algorithmsmentioned above in this section.”
Online Learning: A Comprehensive Survey (2018, methodology): “The micro-clustering algorithms aim to summarize a data in an online manner, and the clustering is performed using these summaries (Cao et al., 2006; Tasoulis et al., 2007; Ruiz et al., 2009; Li-xiong et al., 2009; Ren and Ma, 2009; Ntoutsi et al., 2012).”
Clustering Data Streams Based on Shared Density between Micro-Clusters (2016, methodology): “To adapt to evolving data streams we use the exponential fading strategy introduced in DenStream [6] and used in many other algorithms.”
Progress in Outlier Detection Techniques: A Survey (2019, methodology): “Some of the commonly encountered difficulties related to the nature of the input data, outlier type, data labels, accuracy, and computational complexity in terms of the CPU time andmemory consumption [6]–[9].”
Density-Based Projected Clustering of Data Streams (2012, methodology): “In its online phase, our algorithm adopts the microcluster structure used in most other streaming algorithms [2], [7] with an adaptation to fit our problem (cf.”
Twitter spammer detection using data stream clustering (2014, methodology): “Core-micro-clusters further differ from the traditional coreobjects of DBSCAN with the inclusion of three new attributes, weight, center, and radius.”
A survey on data stream clustering and classification (2015, methodology): “Minimum distance and mean distance are more popular for data streams, for example, D-Stream [24], Den-Stream [22], and MR-Stream [97] use minimum distance for cluster merging, while STREAM [75], ClusStream [3] use mean distance for cluster merging.”
Data clustering: 50 years beyond K-means (2008, neutral): “Because of the high-speed processing requirements, many of the data stream clustering methods (Guha et al., 2003a; Aggarwal et al., 2003; Cao et al., 2006; Hore et al., 2009b) are extensions of simple algorithms such as K-means, Kmedoid, fuzzy c-means, or density-based clustering, modified to work…”
Density-based Projected Clustering over High Dimensional Data Streams (2012, methodology): “A usual way to overcome this problem is by summarizing the data through an appropriate summary structure e.g., [1,2,7,8].”
The ClusTree: indexing micro-clusters for anytime stream mining (2011, methodology): “4.3 by comparing our results against CluStream [2] and DenStream [ 8 ].”
Learning in Nonstationary Environments: A Survey (2015, methodology): “In unsupervised learning/clustering, the learning is done without using labeled data [114]–[116], whereas in active learning (not to be confused with active approaches to learning in nonstationary environments, discussed in Section III-A) the algorithm identifies the most important instances for the”
Introduction to stream: An Extensible Framework for Data Stream Clustering Research with R (2017, methodology): “DSC_DenStream (streamMOA) interfaces MOA’s implementation of the DenStream algorithm by Cao et al. (2006).”