广义神经网络在数据流量预测中的应用仿真
Application Simulation of Generalized Neural Network in Data Traffic Prediction
姚迎乐 1冯乃勤2
作者信息
- 1. 郑州工业应用技术学院软件学院,河南 新郑 451150
- 2. 郑州工业应用技术学院软件学院,河南 新郑 451150;河南师范大学计算机与信息工程学院,河南 新乡 453000
- 折叠
摘要
在超大数据量的流量预测中,数据易出现缺陷、错误或不完整问题.广义神经网络因其对数据处理具有较强的鲁棒性和容错能力,因此研究广义神经网络在数据流量预测中的应用,并对应用效果完成验证.将数据流量预测作为研究目标,分析数据流量特征,获取数据流量时空特征和空间维度两者对应的空间相关性特征.选择和被测网络相关性最大的数据流量作为广义神经网络的输入,构建基于广义神经网络的数据流量预测模型.为验证广义神经网络的应用效果,设计对比测试实验.结果表明,广义神经网络在数据流量预测中具有可行性,且算法应用下数据流量预测误差更小.
Abstract
In the traffic prediction of massive data,data are prone to defects,error or incomplete problems.And the generalized neural network has strong robustness and fault-tolerant ability for processing data.In this paper,data traffic prediction was taken as the research objective,and the characteristics of data traffic were analyzed to obtain the spatial characteristics corresponding to the spatial-temporal characteristics and spatial dimensions of data traffic.Mo-reover,the data flow that correlated the most with the tested network was selected as the input of the generalized neu-ral network.Finally,a model of predicting data flow was built based on the generalized neural network.In order to verify the application effect of generalized neural network,a comparative test was designed.The simulation results show that the generalized neural network is feasible in predicting data traffic,with lower prediction error of data traffic.
关键词
广义神经网络/数据流量/时空特征/空间维度/流量预测Key words
Generalized neural network/Data flow/Spatial-temporal features/Spatial dimension/Flow forecast引用本文复制引用
基金项目
产学合作协同育人项目(第一批)(2021)(202101101016)
省级教学改革项目(2021)(2021SJGLX616)
教育部产学研协同育人项目(220906517024221)
出版年
2024
NETL
NSTL
万方数据