基于神经网络的公路网规模预测

路网规模研究是公路网规划的重要内容。考虑影响公路网合理规模的多种因素，提出了一种基于BP神经网络的公路网规模预测方法，并建立了模拟路网规模与其影响因素间的非线形关系预测模型。步骤依次为：改进传统的BP算法、合理确定影响因素、建立预测模型、模型的训练与检验、数据预测。预测结果表明，该方法客观、合理，预测精度高，实用性强，具有较强的理论与实际应用价值。

Highway Network Scale Prediction Based on BP Neural Network

The research on the scale of highway network is an important part of highway network plan. This paper takes fully account of all kinds of influencing factors, puts forward a prediction technique of highway network scale based on BP neural network, and establishes a prediction model that simulates the nonlinear relation between highway network scale and its influencing factors. The prediction steps are as below, the improvement on BP traditional arithmetic, the reasonable choice of influencing factors, the establishment of prediction model, the training and test of this model, the data of prediction. The prediction result is impartial, rational, accurate and applicable. 2 tabs, 4 figs, 7 refs.