一种数据融合BP神经网络的船舶流量预测算法  

A data fusion BP neural network of ship traffic prediction algorithm

作　　者：刘小东[1] 黄洪琼[1] 

机构地区：[1]上海海事大学信息工程学院,上海201306

出　　处：《微型机与应用》2014年第16期63-66,共4页Microcomputer & Its Applications

基　　金：上海海事大学学术新人培育计划(GK2013087)

摘　　要：为解决现有船舶交通流量预测算法中存在的预测精度不高、算法稳定性差等不足,将一种数据融合BP神经网络的算法用于船舶流量的预测,不仅能较好地实现船舶流量的高精度预测,而且还增强了算法的稳定性。以宁波港口2012年船舶流量观察数据为实例进行分析,用MATLAB软件编程进行系统仿真,实验结果表明,经过数据融合的BP神经网络预测精度高,系统鲁棒性强,预测效果明显优于传统的BP神经网络算法。The existing ship traffic prediction algorithms has the shortages that prediction accuracy is not high , and stability is poor . To solve these problems , this paper proposes a kind of data fusion BP neural network prediction algorithm for ship traffic . It not only can better achieve high accuracy of prediction , but also enhance the stability of the algorithm . Taking Ningbo port 2012 ship traffic observation data as an example for analysis , and using MATLAB software to program . System simulation experiments show that the prediction accuracy is higher after data fusion BP neural network , and system robustness is better than traditional BP neural network algorithm .

关 键 词：数据融合 神经网络 船舶流量 预测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]