一种基于机器学习的5G网络指标处理方法  

A Processing Method of 5G Network Index Based on Machine Learning

作　　者：曹桓 周慧芳 伍岳[2] 许勇 CAO Huan;ZHOU Huifang;WU Yue;XU Yong(China Mobile Design Institute Co.,Ltd.,Sichuan Branch 610045,China;China Mobile Communications Group Sichuan Co.,Ltd.,610041,China)

机构地区：[1]中国移动通信集团设计院有限公司四川分公司,四川成都610045 [2]中国移动通信集团四川有限公司,四川成都610041

出　　处：《通信与信息技术》2022年第S01期19-22,共4页Communication & Information Technology

摘　　要：提出一种5G网络指标处理方法,能够按评估目标自动完成所需评估网络指标的生成与选取,并进一步获得所选网络指标的网络特征权重、获取能够表示网络特征方向的影响系数;该方法可以用于5G网络评估,尤其针对多种类型的网络特征对网络进行多目标、更因分析式的定量评估,能够提升通信运营商网络建设规划前评估寻找问题及问题背后原因的效果,进而提升网络建设资源投向的精准度。This paper presents a 5G network index processing method.This method can automatically generate and select the required evaluation network indicators according to the evaluation objectives.Further,the method can also obtain the weight of the network indicator and obtain the influence coefficient of the network feature,which can indicate the importance of the network feature of the network indicator.This method can be used for 5G network evaluation,especially for various types of networks,such as multi-objective evaluation,and causal analysis evaluation.This method can improve the effect of cause analysis of evaluation before network construction planning of communication operators,and further improve the accuracy of network construction resources.

关 键 词：网络指标 权重确定 机器学习 资源投向精准度 

分 类 号：TN915.63[电子电信—通信与信息系统]