An Improved K-means Text Clustering Algorithm by Optimizing Initial Cluster Centers

Published 1 November 2016

Caiquan Xiong, Zhen Hua, Ke Lv, Xuan Li

DOI10.1109/ccbd.2016.059

Citations64

TL;DR

The experimental results show that the improved K-means algorithm can improve the stability and accuracy of text clustering.

Abstract

K-means clustering algorithm is an influential algorithm in data mining. The traditional K-means algorithm has sensitivity to the initial cluster centers, leading to the result of clustering depends on the initial centers excessively. In order to overcome this shortcoming, this paper proposes an improved K-means text clustering algorithm by optimizing initial cluster centers. The algorithm first calculates the density of each data object in the data set, and then judge which data object is an isolated point. After removing all of isolated points, a set of data objects with high density is obtained. Afterwards, chooses k high density data objects as the initial cluster centers, where the distance between the data objects is the largest. The experimental results show that the improved K-means algorithm can improve the stability and accuracy of text clustering.

Keywords

Computer Science

Munich Personal RePEc Archive (Ludwig Maximilian University of Munich)Some methods for classification and analysis of multivariate observations

22,787 Citations1967James B. MacQueen

Pattern Recognition LettersCluster center initialization algorithm for K-means clustering

678 Citations2004Shehroz S. Khan, Amir Ahmad

An algorithm to compute initial cluster centers for K-means clustering based on two observations that some of the patterns are very similar to each other and that is why they have same cluster membership irrespective to the choice of initial cluster center.

Jisuanji gongchengK-means Clustering Algorithm with Meliorated Initial Center

32 Citations2007Song Xin

Experiments on the standard database UCI show that the proposed method can produce a high purity clustering result and eliminate the sensitivity to the initial start centers.

Computer Engineering and Applications JournalOptimization to K-means initial cluster centers

23 Citations2013Bo Feng, Wenning Hao +2 more

Improved K-means algorithms are proposed that can enhance the clustering performance, get stable clustering in a higher accuracy and solve the problems that the traditional K-Means algorithm has sensitivity to the initial cluster centers.

View PDF

Computer Engineering and Applications JournalImproved k-means initial clustering center selection algorithm

21 Citations2010Hao Zhi-qiang

The clustering results demonstrate that the improved algorithm can enhance the clustering stability and accuracy of ordinary k-means algorithm relatively.

International journal of innovative computing, information & controlAn intelligent initialization method for the K-means clustering algorithm

3 Citations2010許志堅, Jyh‐Jian Sheu +3 more

A two-phase clustering method is proposed by modifying the initialization of K-Means algorithm, which can accomplish the following jobs simultaneously: deciding the proper cluster number automatically, choosing the better initial clusters, and reducing the influence of outliers upon the result of clustering.