统计方法在移动通信网络优化管理的应用

Statistic in the Mobile Communications Network Management Application

当前移动通信网络优化由于设备、时间、空间、工程经验和管理水平方面的差异,存在不同的运行质量评估体系;移动通信网优数据存在着数据量大和分布复杂的特性;并且在管理时必须考虑各地区的移动通信网络发展水平。为了使网络运营质量管理向精细化发展,需要对管理设计给出决策支持。首先,应用基于稳健统计量的盒须图模型,对话务统计指标异常值判定门限进行设定,该方法符合数理统计规律,并且经实验证明合理数据占比大,有一定指导意义,并且提高了效率。其次是应用K均值聚类和Sil houette值,探索地市管理分类,实验说明,此方法具有相当的参考价值。

At present,mobile communication network optimization due to equipment,time,space,engineering experience and manage ment level differences,there are different quality assessment system.Mobile communication network have large data and its distrbution is unknown.It must be considered in management that the district development level of the mobile communication network.In order to man age the quality of mobile communication network and give the decision support for management design,the statistics method must be used.In this paper,first,using box-plot model based on robust estimation to set traffic statistics KPI’s outliers threshold.Experiments show that there is a certain significance and effective.second,applying K-means clustering and Silhouette values,to explore the cities of classification.this method has considerable reference.