摘要
随着5G网络的完善和5G流量的快速增长,如何满足网络连接的多样化需求,提升客户感知保障能力,成为网络运营智能化的关键点。介绍了某省联通在无线网络智能化、潜在贬损用户识别与感知提升方面所做的探索与实践,通过整合XDR话单和专业网管等多源数据,引入熵权法、分层二元评分法等科学算法模型,结合用户体验回访结果不断迭代优化算法,实现了对潜在贬损用户的有效识别,同时在感知驱动网络问题闭环解决方面也取得了较好的效果。
With the improvement of the 5G network and rapid growth of 5G traffic,how to meet diversified network connection requirements and improve customer experience assurance capabilities becomes the key to intelligent network operation.It introduces the exploration and practice of China Unicom in wireless network intelligence,potential derogatory user identification and perception improvement in a certain province.By integrating multi-source data such as XDR and professional NMS data,introducing scientific algorithm models such as entropy weight method,stratified binary scoring method,and continuously optimizing algorithms based on user experience survey results,it can effectively identify potential derogatory users and achieve good results in perception-driven closed-loop resolution of network problems.
作者
毕强
吴彦涛
张冠楠
Bi Qiang;Wu Yantao;Zhang Guannan(China Unicom Jilin Branch,Changchun 130000,China)
出处
《邮电设计技术》
2024年第5期14-18,共5页
Designing Techniques of Posts and Telecommunications
关键词
无线智能化
大数据
贬损用户识别
Wireless intelligence
Big data
Derogatory user identification
