船舶航行网络流量的建模与预测

Modeling and forecasting of navigation network traffic

船舶航行网络流量受到外界因素的干扰,具有比较强的随机性,当前船舶航行网络流量的预测准确性差,为了改善船舶航行网络流量预测的效果,设计一种高精度的船舶航行网络流量建模与预测方法。首先收集一维船舶航行网络流量样本,并通过变换得到一种多维的船舶航行网络流量样本,然后引入极限学习机描述船舶航行网络流量的变化规律,并对极限学习机参数进行优化,改进基本极限学习机的不足,最后进行船舶航行网络流量预测的应用实例,分析本文方法的可行性,结果表明,本文方法的船舶航行网络流量预测误差小于5%,低于实际应用要求的10%,同时船舶航行网络流量建模过程自化程度高,简单,获得了较快的船舶航行网络流量预测速度,为解决船舶航行网络流量的预测问题提供了一种建模技术。

The ship navigation network traffic is interfered by external factors,has a strong randomness,and the current ship navigation network traffic prediction accuracy.In order to improve the ship navigation network traffic prediction effect,a high-precision ship navigation network flow is designed.Modeling and prediction methods.Firstly,the traffic samples of the one-dimensional ship navigation network are collected,and a multi-dimensional ship navigation network traffic sample is obtained by transformation.Then,the limit learning machine is introduced to describe the variation law of the ship navigation network flow,and the parameters of the extreme learning machine are optimized to improve the basic limit.Insufficient learning machine,the application example of ship navigation network traffic prediction is carried out,and the feasibility of this method is analyzed.The results show that the ship navigation network traffic prediction error of this method is less than 5%,which is lower than the actual application requirement of 10%.The ship navigation network traffic modeling process is highly self-contained and simple,and obtains a faster ship navigation network traffic prediction speed,which provides a modeling technique for solving the ship navigation network traffic prediction problem.