多跳网络中分类属性数据模糊聚类仿真
邓峰
【期刊名称】《计算机仿真》 【年(卷),期】2017(034)001
【摘要】针对多跳网络中分类属性数据进行聚类,为提高多跳网络数据管理效率.多跳网络分类属性数据聚类时,需要对孤立数据和异常数据进行单独聚类划分,但是传统的距离差异度组合权重算法没有孤立数据和异常数据检测规则,不能很好的识别两类数据,导致分类属性数据聚类效果差.提出一种多跳网络中分类数据模糊聚类的方法.针对传统聚类方法产生的弊端进行分析,利用多跳网络的特殊属性,对多跳网络中模糊属性数据建立模糊等价关系,对多属性数据实现规格化处理得到任意属性值从属于某种共同的数据特性范围,最终实现聚类.仿真结果表明,在多跳网络系统分类属性数据聚类中,改进算法具有较高的聚类准确性,聚类性能较好.%It can improve management efficiency of multi-hop network data to cluster its classification attribute data.It needs to cluster and divide isolated data and anomalous data separately during clustering classification attribute data.However,traditional combination weight algorithm of distance discrepancy degree does not have detection rule of isolated data and anomalous data.It cannot identify two kinds of data well.So it results in poor clustering effect of classification attribute data.In this paper,a fuzzy clustering method of classification data in multi-hop network is proposed.The disadvantage of traditional clustering method is analyzed.Special property of multi-hop network is