您当前访问的的浏览器版本过低,为了给您带来更好的体验,建议您升级至Edge浏览器或者推荐使用Google浏览器
取消

基于机器学习的无线网络流量预测与增长潜力评估

作者:邢旭东,高晖,顾军 阅读量:1484

基于机器学习的无线网络流量预测与增长潜力评估

邢旭东1,高晖1,顾军2
(1. 北京邮电大学可信分布式计算与服务教育部重点实验室,中国 北京 100876; 2. 中兴通讯股份有限公司,中国 深圳 518057 )

摘要:提出一个基于机器学习的无线网络流量预测及流量增长潜力评估方案。该方案分析蜂窝网络中的实际业务流量数据在时间维度上的变化规律,并借助高斯过程的机器学习方法来预测业务变化趋势,从短期角度为运营商的网络优化部署提供指导。基于极限梯度提升(XGBoost)机器学习框架,建立网络中其他运营数据与业务流量的多维映射关系,应用改进的量子粒子群算法进一步寻找蜂窝小区所能承载的流量上限,从长期角度为网络优化部署提供指导,提升网络流量水平、释放流量增长潜力。  
关键词:机器学习;移动网络数据分析;流量预测;流量增长潜力评估  


Traffic Prediction and Growth Potential Evaluation in Wireless Network Based on Machine Learning

XING Xudong1, GAO Hui1, GU Jun2
(1. Key Laboratory of Trustworthy Distributed Computing and Service, Beijing University of Posts and Telecommunications, Ministry of Education, Beijing 100876, China; 2. ZTE Corporation, Shenzhen 518057, China )

Abstract: A wireless network traffic prediction and traffic growth potential evaluation scheme based on machine learning is proposed. Based on the actual traffic data in the cellular network, this scheme analyzes the change rule in the time dimension and uses the machine learning method of Gaussian process to predict the trend of traffic, which provides guidance for the network optimization deployment of operators in the short term. Based on the eXtreme Gradient Boosting (XGBoost) machine learning framework, the multi-dimensional mapping relationship between other operation data and traffic in the network is established, and the revised quantum particle swarm optimization algorithm is applied to further find the upper limit of traffic that the cellular cell can carry, so as to provide guidance for network optimization deployment from a long-term perspective, improve the network traffic level, and release the traffic growth potential.  
Keywords: machine learning; mobile network data analysis; traffic prediction; traffic growth potential evaluation

在线PDF浏览:PDF