基于BP神经网络的运量预测模型优化研究

在传统多种单项预测模型与组合预测方法的基础上,利用BP神经网络技术的非线性映射能力,在多个预测模型与实际数列之间建立一种非线性关系,对运量预测结果进行优化,以达到提高预测精度的目的.通过实例分析,表明这种经过BP神经网络优化后的预测模型,可一定程度上克服传统单个预测模型的部分局限性,提高预测精度,用于运量预测是可行的.

Optimization of Traffic Volume Prediction Model Based on BP Neural Network

In order to optimize the prediction results and raise forecast accuracy, this paper established nonlinear relationship between prediction model and practical sequences by using the nonlinear mapping ability of BP neural network based on traditional single prediction models and combination of forecasting method. The case study showed that the prediction model optimized by BP neural network can overcome some limitations of traditional single prediction models, and raise forecast accuracy. This method based on BP neural network used in traffic volume prediction is feasible.