摘要
5G通信是现代信息体系的重要基础,由于流量波动较大,建立有效的预测与分配机制对于降低基站负载、提高信息传输效能具有积极意义。文中优化了ER-LSTM预测算法,采用误差循环的方式引入机器学习架构,旨在提高预测精度,并形成有效的分配机制。研究发现,优化后的算法模型在500~2500次循环学习的基础上能实现有效的预测,提高系统承载力,具有推广应用的价值。
5G communication is an important foundation of modern information system.Due to the large fluctuation of traffic,establishing an effective prediction and allocation mechanism is of positive significance for reducing the load of base stations and improving the efficiency of information transmission.In this paper,the ER-LSTM prediction algorithm is optimized,and the machine learning architecture is introduced by using the error cycle method,aiming to improve the prediction accuracy and form an effective allocation mechanism.The study found that the optimized algorithm model can achieve effective prediction on the basis of 500-2500 cycles of learning,improve the system carrying capacity,and has the value of popularization and application。
作者
代永亮
DAI YongLiang(Yinan County Disabled Persons Federation,Linyi,Shandong 276300,China)
出处
《移动信息》
2024年第6期17-19,22,共4页
MOBILE INFORMATION
关键词
机器学习
5G
流量预测
切片资源分配
Machine learning
5G
Traffic prediction
Slice resource allocation
作者简介
代永亮(1983-),本科,中级网络工程师,研究方向为电子信息。
