因果关系模型与智能预测技术在广电主干光缆传输网中的研究与实践  

Research and Practice of Causal Relationship Model and Intelligent Prediction Technology in the Backbone Optical Cable Transmission Network of Radio and Television

作　　者：施小明 Shi Xiaoming(Zhejiang Guanglian Cable Television Transmission Center,Zhejiang 310012,China)

机构地区：[1]浙江广联有线电视传输中心,浙江310012

出　　处：《广播与电视技术》2024年第4期82-87,共6页Radio & TV Broadcast Engineering

摘　　要：本文基于新型因果关系模型和复杂网络理论,构建了省级广电主干网光链路因果关系与拓扑结构,并重点对主干网光链路海量历史数据进行模型训练,挖掘光链路数据特征,实现了高质量光链路趋势性预测。经过仿真测试以及在省主干网数智化综合运维平台中的部署应用,证明因果关系模型及人工智能技术可以很好地应用于主干网光链路的时间序列大数据场景中,有效提升了主干网的运维效率及预警能力,具有较高的推广价值。Based on a novel causal relationship model and complex network theory,this paper establishes the causal relationships and topological structure of optical links in the provincial-level broadcasting and television backbone network.It focuses on extensive model training using historical data from the backbone network's optical links,extracting features from the optical link data,and proposes a trend prediction technology for optical links.Through simulation testing and deployment in the province's backbone network digitalized comprehensive operation and maintenance platform,the causal relationship model and artificial intelligence technologies have proven to be effective in handling time-series big data scenarios for backbone network optical links,significantly enhancing the operational efficiency and warning capabilities of the backbone network,and demonstrating high potential for widespread application.

关 键 词：智慧广电 光链路 滑动窗口 人工智能 长短时记忆网络 

分 类 号：TN929.11[电子电信—通信与信息系统]