The global k-means clustering algorithm

By Aristidis C. Likas, Nikos A. Vlassis, Jakob Verbeek. Pattern Recognition, 2003

3,390 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Signal Processing

View Paper: https://www.semanticscholar.org/paper/c82e5c28fb271280a628588663b9eb66b262b6b0

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. Pasi Fränti
2. Sami Sieranoja
3. Ranjan Maitra
4. Volodymyr Melnykov
5. Jakob Verbeek
6. Ben J. A. Kröse
7. Yang Liu
8. Stefano Zanero

In Signal Processing

1. Pasi Fränti
2. Sami Sieranoja
3. Jakob Verbeek
4. Ben J. A. Kröse
5. Stefano Zanero
6. Ranjan Maitra
7. Olli Virmajoki
8. Sergio M. Savaresi

Citations in the corpus

• DialogSum: A Real-Life Scenario Dialogue Summarization Dataset (2021, methodology): “We use k -means (Likas et al., 2003) to cluster the dialogue topic with GloVe embedding (Pennington et al., 2014), where k = 20 .”
• Machine learning methods for turbulence modeling in subsonic flows around airfoils (2018, background): “There are various clustering methods and the distance calculating formula is dependent on the specific problem [37-40].”
• A Comparative Study of Efficient Initialization Methods for the K-Means Clustering Algorithm (2012, methodology): “The global k-means method [53] takes ¯ X as the first center.”
• Efficient Greedy Learning of Gaussian Mixture Models (2003, methodology): “We recently applied a similar strategy to the k-means clustering problem (Likas et al., 2002), for which encouraging results were obtained.”
• Survey of clustering algorithms (2005, background): “Likas, Vlassis, and Verbeek proposed a global -means algorithm consisting of a series of -means clustering procedures with the number of clusters varying from 1 to [186].”
• Mechanical behaviour and damage mechanisms analysis of a flax-fibre reinforced composite by acoustic emission (2016, positive): “The K-mean algorithm [29] was used for the unsupervised pattern recognition.”
• Balanced K-Means for Clustering (2014, background): “To get better results than in k-means, slower agglomerative algorithms [10,13,12] or more complex k-means variants [3,11,21,18] are sometimes used.”
• Unsupervised learning techniques for an intrusion detection system (2004, methodology): “K-means is globally the fastest algorithm we tested, showing really no performance problem; however, the corrections necessary to eliminate or reduce the random initialization weakness (for instance using the so-called “global K-means” algorithm [23]) make the algorithm completely untractable.”
• Deep learning and machine vision for food processing: A survey (2021, background): “Unlike the K-means algorithm (which clusters each sample into a certain class and assigns a label to the sample), fuzzy clustering assigns a probability vector (rather than a label) to a sample (Ghosh et al., 2013).”
• How much can k-means be improved by using better initialization and repeats? (2019, background): “Steinley and Brusco [21] studied 12 variants including complete algorithms like agglomerative clustering [27] and global k-means [28].”
• A Comprehensive Deep Learning-Based Approach to Reduced Order Modeling of Nonlinear Time-Dependent Parametrized PDEs (2020, methodology): “In particular, we employ the k-means clustering algorithm [38], an unsupervised statistical learning technique for finding clusters and cluster centers in an unlabeled dataset, to partition into Nc clusters the snapshots, i.”