初始点优化与参数自适应的密度聚类算法  

Density Clustering Algorithm with Initial Point Optimization and Parameter Self-adaption

作　　者：戴阳阳 李朝锋[1] 徐华[1] 

机构地区：[1]江南大学物联网工程学院,江苏无锡214122

出　　处：《计算机工程》2016年第1期203-209,共7页Computer Engineering

基　　金：国家留学基金资助项目(201308320030);江苏省自然科学基金资助项目(BK20140165)

摘　　要：针对密度聚类算法DBSCAN无法处理变化密度的问题,提出一种初始点优化与参数自适应的改进算法。利用初始点优化方法确定全局密度最大的点,结合该点和数据集自身的特征,自适应得到DBSCAN算法聚类出当前簇所需要的合适参数。该算法能够为不同密度的簇自适应设置不同的参数,而且优先对高密度簇进行聚类,即能对变化密度的数据集进行聚类。实验结果表明,该算法可以发现任意形状、大小和变化密度的簇,解决数据重叠和簇内密度不均匀问题,具有较高的聚类准确率。Aiming at the problem that the Density Based Spatial Clustering algorithm of Application with Noise（DBSCAN） can not find clusters of varied densities,this paper proposes a density clustering algorithm with initial points optimization and parameter self-adaption.It uses the method of optimization initial points to find the maximum density point in the current global datasets,and adaptively calculates the parameters of DBSCAN for the current cluster with the features of the current maximum density point and current datasets.These parameters are found to be different with the other clusters＇ parameters,and the high-density cluster gets priority processed,so this algorithm can find clusters of varied density.Experimental results demonstrate that the improved algorithm can find clusters of arbitrary shape,size and density,enhance the ability to deal with overlapping data and uneven density in the cluster,and get clustering in higher accuracy.

关 键 词：初始点优化 自适应 变化密度 聚类 数据挖掘 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]