基于联系数的位置不确定性数据UCNK-Means聚类算法

摘要位置不确定性数据的聚类是一个新的不确定性数据聚类问题。其聚类方法主要包括获取对象的概率密度函数,通过积分计算对象间的期望距离来进行聚类分析和以区间数表示对象,通过区间数的系列运算来进行聚类分析这两大类。前者存在概率密度函数获取困难、计算复杂、实用性不强的缺陷；后者在区间数转化为实数过程中,忽略了区间数变化范围对聚类效果的影响,其聚类质量不佳。鉴于此,提出一种基于联系数的不确定对象聚类新算法UCNK-Means。该算法用联系数巧妙地表示不确定性对象,并专门定义了对象间的联系距离,运用联系数态势值比较联系距离大小,克服了现有算法的不足。仿真实验表明,UCNK-Means具有聚类精度高、计算复杂度低、实用性强的特点。

UCNK-Means Clustering Method for Position Uncertain Data Based on Connection Number

Clustering for position uncertain data is a new problem of uncertain data clustering.Mainly there are two solutions to this new problem.The first is clustering acquiring the probability density function or probability distribution function of uncertain object and getting the expected distance with integral operation.The second is clustering by series of operation of interval data.However,the former has the disadvantages of getting probability density function hard,complex operation and poor practicability,and the latter ignores the effect of the range of interval data to the result of clustering.Therefore,a new uncertain data clustering method UCNK-Means was put forward.This method uses connection number as the model of uncertain object and defines connection distance between two objects and uses the situationvalue to compare the connection distance,which overcome the disadvantages existed in the two solutions above.The experiment illustrates that UCNK-Means has high precision of clustering,low complexity and strong praticability.