Extended K-Means Algorithm
Authors:
- Faliu YiFYFaliu Yi
- Inkyu MoonIMInkyu Moon
- 1 August 2013
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 2, 263-266
Abstract
In the conventional K-means algorithm, the input data are automatically grouped into corresponding cluster by minimizing the within-cluster sum of squares. However, the traditional K-means algorithm doesn't do any constraints to the number of elements in each group. In the area of logistics management, each cluster will need to satisfy with a predefined number of elements. Thus, the clustering algorithm with controlled number of elements in each group is necessary. In this paper, we present a new method called extended k-means algorithm to extend the ordinary K-means approach. In this approach, the number of element in each group is adjusted by using greedy algorithm and the experimental results show that this extended K-means algorithm can work well for grouping data where the numbers of elements in each group need to be restrained.Keywords
This publication has 4 references indexed in Scilit:
- A fast k-means implementation using coresets2006
- An efficient k-means clustering algorithm: analysis and implementationIEEE Transactions on Pattern Analysis and Machine Intelligence, 2002
- Least squares quantization in PCMIEEE Transactions on Information Theory, 1982
- Algorithm AS 136: A K-Means Clustering AlgorithmJournal of the Royal Statistical Society Series C: Applied Statistics, 1979