一种改进的聚类和孤立点检测算法

对基于距离的聚类及基于密度的孤立点检测方法进行了分析研究,提出了一种基于距离和密度的聚类和孤立点检测算法DDBCOD.该算法根据距离和密度阈值对数据进行聚类,并发现数据中的孤立点.实验表明,该算法能够识别任意形状的聚类,对高维数据有效,能够很好地识别出孤立点.

A Improved Clustering and Outlier Detection Algorithm

In this paper,the author briefly analyzes clustering and outlier detection methods. A distance&density-based clustering and outlier detection algorithm (DDBCOD) is proposed in,it records the datum points by distance and density threshold, and identifies outliers by density threshold. As shown in the experimental results,the DDBCOD algorithm can cluster the dataset properly, it can discover clusters of arbitrary shapes, it is valid for high dimension dataset,and it can find outliers accurately and validly.