| 运用前向神经网络法预测铁路客运量 |
| |
| 引用本文： | 覃频频,晏启鹏,黄大明. 运用前向神经网络法预测铁路客运量[J]. 铁道运输与经济, 2005, 27(12): 95-98 |
| |
| 作者姓名： | 覃频频 晏启鹏 黄大明 |
| |
| 作者单位： | 1. 西南交通大学,交通运输学院,四川,成都,610031
2. 广西大学,机械工程学院,广西,南宁,530004 |
| |
| 摘 要： | 为提高铁路客运量的预测精度,应用一种非线性预测方法:多层前向神经网络建立铁路客运量预测模型。在介绍误差修正学习算法和误差反向算法的基础上,通过预测实例计算,与其他3个常用预测模型:多元回归模型、简单移动模型和平均移动模型进行预测比较,结果表明误差反向算法的多层前向神经网络模型预测精度最高。
|
| 关 键 词： | 铁路 客运量 预测 误差 前向神经网络 |
| 文章编号： | 1003-1421(2005)12-0095-03 |
| 收稿时间： | 2005-04-01 |
| 修稿时间： | 2005-08-18 |
Railway Passenger Volume Forecast by Means of Feed-Forward Neutral Network Method |
| |
| QIN Pin-pin,YAN Qi-peng,HUANG Da-ming. Railway Passenger Volume Forecast by Means of Feed-Forward Neutral Network Method[J]. Rail Way Transport and Economy, 2005, 27(12): 95-98 |
| |
| Authors: | QIN Pin-pin YAN Qi-peng HUANG Da-ming |
| |
| Affiliation: | 1.School of Traffic and Transportation, South-West Jiaotong University, Chengdu, Sichuan 610031, China; 2.College of Mechanical Engineering, Guangxi University, Guangxi, Nanning 530004, China |
| |
| Abstract: | To improve the accuracy on forecast of railway passenger volume,a non-linear forecast method,multi-feed-forward neural network,is adopted asforecasting model. Based on introducing discrepancy correction algorithm anddiscrepancy back-propagation algorithm,this paper carries out practicalcalculation of forecast and compares the result with that of other three modelsoften utilized: multi-regression model,SMM and AMM. The comparison showsthat the discrepancy back-propagation algorithm of multi-feed-forward neuralnetwork model has the best accuracy in forecasting. |
| |
| Keywords: | railway passenger volume forecasting discrepancy feed-forwardneural network |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
