基于PSO-RBF组合模型的长江集装箱运价指数预测
Prediction of Yangtze River Containerized Freight Index Based on PSO-RBF Combined Model
黄建华 1缪思琪1
作者信息
- 1. 福州大学 经济与管理学院,福建 福州 350108
- 折叠
摘要
长江集装箱运价指数作为长江航运市场的晴雨指向标,能够有效反映中国长江航运的经济情况,同时侧面反映出中国内河航运的发展动态.通过对长江集装箱运价指数的预测,可以为沿岸航运企业经营决策和政府宏观经济制定提供重要依据.选取影响长江集装箱运价指数的 8 个指数,运用BP神经网路、RBF神经网络对2017 年至2022 年5 月长江集装箱运价指数进行预测,提出了一种改进的PSO-RBF组合模型,获得的预测误差较小.结果表明:粒子群算法能对RBF神经网络的输出权重、隐单元中心等关键参数取值进行寻优,使其能够更好地收敛,结果优于其他算法;PSO-RBF组合模型是预测长江集装箱运价指数的一种有效方法.
Abstract
Yangtze River containerized freight index,as a barometer of the Yangtze River shipping market,can effectively reflect the economic situation of China's Yangtze River shipping and reflect the development dynamics of China′s inland waterway shipping.The forecast of Yangtze River containerized freight index can provide important basis for the management decision of coastal shipping enterprises and the government′s macroeconomic formulation.Eight indexes affecting the Yangtze River contain-erized freight index were selected,BP neural network and RBF neural network were used to forecast the Yangtze River container-ized freight index from 2017 to May 2022,and an improved PSO-RBF combination model was proposed,which obtained a small prediction error.The results showed that the particle swarm optimization algorithm can optimize the key parameters of the RBF neural network,such as the output weight and the hidden cell center,so that the RBF neural network can converge better,and the results are better than other algorithms.The results showed that PSO-RBF combination model is an effective method to pre-dict the Yangtze River containerized freight index.
关键词
长江集装箱运价指数/粒子群算法/RBF神经网络/组合模型/预测Key words
Yangtze River containerized freight index/particle swarm optimization/RBF neural network/combination mod-el/prediction引用本文复制引用
出版年
2024
NETL
NSTL
维普
万方数据