中文摘要：

在网络化测试测量信息体系的不确定性测量数据聚类方法研究中,普遍假定测量数据的概率密度函数或者概率分布函数等信息是已知的,这与实际应用系统中这些信息难以获取的情况是相悖的,鉴于此,利用区间数的方法,结合测量数据的统计值来合理地表示多维不确定性测试测量数据,并采用低计算复杂度的不确定性数据距离计算方法,提出一种基于区间数的多维不确定性数据聚类方法——UIDK-means。实验结果表明,该方法具有较高的聚类精度和较低的计算复杂度。

英文摘要：

In uncertain measurement data clustering methods for networked measurement and test information system,most methods assume the probability density function or probability distribution function of the measurement data is known,which is in contradiction with the issue that this information is rarely available.So in this paper,interval data combined with statistic information is used to express multi-dimensional uncertain measurement data reasonably,a new uncertain distance computing method is proposed to measure the similarity of different uncertain data.And a new uncertain multi-dimension data clustering algorithm—UIDK-means based on the interval data is proposed and applied to uncertain measurement data.Experiment results show that the uncertain clustering algorithm can obtain better clustering precision with low computing complexity.