基于神经网络算法的感知问题定位方法OA
Perception problem localization method based on neural network algorithm
随着4G/5G网络的规模化部署与用户对网络服务质量要求的不断提升,传统基于人工经验的网络感知问题定位方法在准确性、效率与全面性方面面临严峻挑战.本文提出一种基于BP神经网络的用户感知问题定位方法,通过构建多层神经网络映射模型,结合欧式距离算法与相关性分析,实现对问题小区的精准定位与根因剖析.该方法包含三个核心步骤:投诉小区定位、小区网络问题神经画像构建、网络问题数据清洗与优先级排序.本方法在四川全省应用中显著提升了问题处理效率与客户满意度,实现了从"人工排查"到"智能诊断"的转变,具备良好的工程应用价值.
With the large-scale deployment of 4G/5G networks and the increasing demands of users for network service quality,tra-ditional network perception problem localization methods based on human experience face severe challenges in terms of accuracy,effi-ciency,and comprehensiveness.This paper proposes a user-perceived problem localization method based on BP neural networks.By con-structing a multi-layer neural network mapping model and combining Euclidean distance algorithm and correlation analysis,it achieves accurate localization and root cause analysis of problematic cells.This method includes three core steps:locating the complained cell,constructing a neural profile of the cell's network problem,and cleaning and prioritizing network problem data.This method has signifi-cantly improved problem handling efficiency and customer satisfaction in Sichuan,realizing a shift from"manual investigation"to"intel-ligent diagnosis,"and possesses significant engineering application value.
简晨;苟浩淞
中国移动(成都)产业研究院,四川 成都 610213中国移动(成都)产业研究院,四川 成都 610213
信息技术与安全科学
5G建设BP神经网络问题定位网络优化
5GBP neural networkProblem localizationNetwork optimization
《通信与信息技术》 2026 (5)
18-21,4
评论