模糊神经网络非线性组合预测在铁路货运量预测中的应用

Application of Nonlinear Combination Forecast of Fuzzy Neural Networkon Forecasting Railway Freight Volume

准确的铁路货运量预测关系到铁路运输的发展,为此提出模糊神经网络非线性组合预测模型,应用三次指数预测模型、灰色理论预测模型、多元回归预测模型的预测值作为模糊神经网络的测试样本数据库,输出样本为铁路货运量,并采用全局优化的粒子群算法优化模糊神经网络的参数。仿真结果表明该模型能够取得比单项预测模型更高的精度。

Accurate forecasting for railway freight volume is important to the development of railway transportation. This paper presents nonlinear combination forecast model of fuzzy neural network, uses the results of three-time exponential forecast model, grey theory forecast model and multiple regression forecast model as the test sample database of fuzzy neural network, thereinto the output sample is the railway freight volume, and also uses the whole optimized particle swarm algorithm to optimize the parameters of fuzzy neural network. The simulation results demonstrate the proposed model can improve the accuracy compared with individual forecasting model.