摘要:
在高维数据聚类中,受维度效应的影响,现有的算法聚类效果不佳。为此,提出一种适用于高维数据的密度聚类算法StaDeCon。在经典的PreDeCon算法基础上,引入子空间维度权重的计算方法,避免PreDeCon算法使用全空间距离度量带来的问题,提高了聚类的质量。在合成数据和实际应用数据集上的实验结果表明,该算法在高维数据聚类上可取得较好的聚类精度,算法是有效可行的。
关键词:
聚类,
高维数据,
子空间,
维度加权
Abstract: In clustering of high dimensional data, most of the existing algorithms can not reach people’s expectation due to the curse of dimensionality. Based on the classic PreDeCon algorithm, this paper presents the StaDeCon, a density clustering algorithm for high dimensional data, which introduces a measure of subspace dimensional weighting to avoid the problem existing in PreDeCon caused by using full dimensional distance, and in this way, the quality of clustering is improved. Experimental results both on artificial and practical data show that the algorithm is more accurate, and it is effective and feasible.
Key words:
clustering,
high dimensional data,
subspace,
dimensional weighting
中图分类号:
黄王非;陈黎飞;姜青山;. 基于子空间维度加权的密度聚类算法[J]. 计算机工程, 2010, 36(9): 65-67.
HUANG Wang-fei; CHEN Li-fei; JIANG Qing-shan;. Density Clustering Algorithm Based on Subspace Dimensional Weighting[J]. Computer Engineering, 2010, 36(9): 65-67.