A method for initialising the K-means clustering algorithm using kd-trees

By Stephen James Redmond, Conor Heneghan. Pattern Recognition Letters, 2007

326 citations recorded by the source corpus.

How researchers see this paper

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

A Comparative Study of Efficient Initialization Methods for the K-Means Clustering Algorithm (2012, methodology): “Redmond and Heneghan’s method [44] ﬁrst constructs a kd-tree of the data points to perform density estimation and then uses a modiﬁed maximin method to select K centers from densely populated leaf buckets.”
Fast approximate spectral clustering (2009, methodology): “We chose to study three initialization methods, based on their documen ted favorable performance [6, 30, 31, 25], as well as their relatively straightforward implementation: the H artigan-Wong algorithm (KM-1) [16], the samplingbased two-stage algorithm (KM-2) (i.”
Medical Image Segmentation Methods, Algorithms, and Applications (2014, methodology): “Some reﬁnements have been done on this algorithm to improve the problem and an enhanced algorithm has been presented [73 (cid:1) 77].”
kml and kml3d: R Packages to Cluster Longitudinal Data (2015, background): “Some authors have therefore proposed various initialization methods (Pena, Lozano, and Larranaga 1999; Khan and Ahmad 2004; Redmond and Heneghan 2007; Steinley and Brusco 2007).”