Data clustering using evidence accumulation

By Ana L. N. Fred, Anil K. Jain. Object recognition supported by user interaction for service robots, 2002

569 citations recorded by the source corpus.

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

View Paper: https://www.semanticscholar.org/paper/e1954d2da335095ec547640612863585b81f16b0

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 Computer Vision and Pattern Recognition

In Signal Processing

Citations in the corpus

A Mixture Model for Clustering Ensembles (2004, methodology): “Further work by Fred and Jain [12] also used co-association values, but instead of a fixed threshold, they applied a hierarchical (single-link) clustering to the co-association matrix.”
Evaluation of Stability of k-Means Cluster Ensembles with Respect to Random Initialization (2006, methodology): “; L) [9], [32], [37], the feature-based approach (treating outputs from the clusterers as L categorical features and building a clusterer thereupon) [35], the hyper-graph approach (constructing a hyper-graph representing the total output from the clusterers and cutting the redundant edges) [32], and”
Advances in Fuzzy Clustering and its Applications (2007, methodology): “We will also present some related work on robust clustering by Fred and Jain (2002) and on generative models for ensembles by Topchy et al. (2004).”
Combining multiple clusterings using evidence accumulation (2005, methodology): “An improvement of the well-known SL method, by exploring nearestneighbor relationships, leads to a more efficient algorithm than the one introduced in [53], [54], [55], and extends the rangeof applicability of theproposed technique to largerdata sets.”
Solving cluster ensemble problems by bipartite graph partitioning (2004, methodology): “Recall that the commonly used agglomerative approach (Fred & Jain, 2002; Fern & Brodley, 2003; Monti et al., 2003) generates a similarity matrix from the cluster ensemble and then performs agglomerative clustering using the similarity matrix.”
Cluster ensembles --- a knowledge reuse framework for combining multiple partitions (2003, methodology): “More recently, an ‘evidence accumulation’ framework was proposed wherein multiple k-means, using a much higher value of k than the final anticipated answer, were run on a common data-set (Fred and Jain, 2002).”
Cluster center initialization algorithm for K-means clustering (2004, methodology): “Khan and Kant (2007) propose a method that is based on the idea of evidence accumulation for combining the results of multiple clusterings (Fred & Jain, 2002) and only focus on those data objects that are less vulnerable to the choice of random selection of modes and to choose the most diverse set…”
Using diversity in cluster ensembles (2004, methodology): “Perhaps the simplest implementation of the generic cluster ensemble algorithm is as follows (called " voting c-means algorithm " in [5] and " evidence accumulation algorithm " in [6] Join in the same cluster all the points whose pairwise entry in M is greater than θ.”
Robust data clustering (2003, methodology): “Instead of choosing a particular clustering algorithm for a given data set, the idea of combining the results of multiple clusterings in order to obtain robust data partitions has recently been proposed [13, 21].”
ClusterFit: Improving Generalization of Visual Representations (2019, background): “Using evidence accumulation meth-ods [17, 41, 50] for clustering is another worthwhile direction to explore.”
DUET: integration of dynamic and static analyses for malware clustering with cluster ensembles (2013, background): “Fed [10] and Jain [11] constructed the consensus function as a co-association matrix which represents the association or connectivity be-tween each pair of data samples.”
Cluster Ensembles --- A Knowledge Reuse Framework for Combining Multiple Partitions (2002, methodology): “The co-occurrence matrix is analogous to the one used for CSPA, except, of course, that the clusterings generated in Fred and Jain (2002) are not legacy and are at a finer level of resolution.”
Clustering ensembles: models of consensus and weak partitions (2005, methodology): “We have investigated the performance of a family of consensus functions based on categorical clustering including the coassociation-based hierarchical methods [15], [16], [17], hypergraph algorithms [47], [29], [30], and our new consensus functions.”
Ensemble learning: A survey (2018, background): “This approach has been shown to improve the stability and the robustness of clustering models (Fred & Jain, 2002) — that is, the resulted clustering systems are less sensitive to outliers and noise, and they usually outperform single clustering models.”
Data clustering: 50 years beyond K-means (2008, unknown): “The success of ensemble methods for supervised learning has motivated the development of ensemble methods for unsupervised learning (Fred and Jain, 2002).”
Combining multiple weak clusterings (2003, methodology): “One can use a co-association matrix for subsequent clustering by single-link algorithm [4] or any other type of agglomerative procedure to obtain a target”