数据挖掘中基于模型的聚类分析方法研究

The Study of Methods of Clustering Analysis Based Model in Dataming

聚类分析已成为数据挖掘研究中非常活跃的研究课题,在聚类分析方法中,基于模型的算法由于考虑到“噪声”或异常数据,可以自动确定聚类个数,可以产生鲁棒的聚类方法,而成为领域研究的一个重点。本文主要对神经网络中的竞争学习神经网络、SOFM 神经网络方法、统计学聚类方法研究。

Clustering analysis has been a active field in datamining ,the paper introduces major Clustering analysis methods based model: probability method, Competitive Cearning NN and SOFM NN, the arithmetics can deal with outline data and noise data, at the same time confirm the numbers of clustering ,at last produce better Methods of Clustering analysis.