基于AI算法的携号转网预警模型搭建  

Construction of MNP early warning model based on AI algorithm

作　　者：黄美思 黄清钧 蒋弢 HUANG Meisi;HUANG Qingjun;JIANG Tao(China Unicom Network Chengdu Branch,Chengdu 610000,China)

机构地区：[1]中国联合网络通信有限公司成都市分公司,四川成都610000

出　　处：《通信与信息技术》2023年第6期108-112,共5页Communication & Information Technology

摘　　要：通过对携号转网用户转网前三个月的网络使用情况、投诉处理情况以及产品应用情况等数据的分析,根据用户携出的主观行为和外部客观特征,找出相关影响因子:是否拨打过客服投诉、语音通话是否有较大波动、套餐资费是否超出、常驻网络是否有问题(使用字典分位法进行判断)等大数据信息。最终使用这些影响因子建立逻辑线性回归算法模型,搭建基于AI算法的携号转网预警模型,并与实际携出情况进行对比,在AI算法训练成功后,能对全网用户的携号转网情况进行预测预警。Through the analysis of the data on network usage,complaint handling and product application of the three months prior to the switching of the number portability users,the relevant influencing factors were identified based on the subjective behavior of the users'portability and external objective characteristics:whether they had dialed the customer service to complain,whether there was a large fluctuation in the voice calls,whether the package tariffs exceeded the tariffs,and whether there was a problem with the resident network(using the dictionary partitioning method to make judgments),and other big data Information.Finally,the logistic linear regres⁃sion algorithm model is established using these influencing factors,and the early warning model based on AI algorithm is built,and compared with the actual carry-out situation,after the successful training of AI algorithm,it can predict and warn the carry-out situa⁃tion of the users in the whole network.

关 键 词：离网 影响因子 字典分位法 逻辑线性回归算法 

分 类 号：TN915.06[电子电信—通信与信息系统]