基于灰色理论与BP神经网络的交通运输量组合预测研究

交通运输量预测是公路规划与管理中的重要问题.结合灰色理论模型与神经网络的优点,基于不同的组合预测思想分别建立了加权算术平均组合、加权平方和平均组合及加权比例平均组合预测模型.利用不同的评价模型对多组数据进行了评价分析研究.结果表明,通过选择合适的组合预测模型及参数估计方法并确定最佳的模型参数,能够有效地提高预测精度.

Fuzzy reasoning method of rank feature value characteristic expanding for forecasting mid-long runoff

Traffic forecasting is an important problem of road planning and management. Gray model and neural network are used for forecasting traffic with their own advantages. Their combining forecasting models with weighted arithmetic average, weighted average square sum and weighted ratio average are proposed to forecast traffic. Multi-group data are analyzed by different estimating models, and results of examples show that the combining model is capable of improving the precision of prediction by selecting suitable models and their parameters.