快速路小时交通量组合预测方法研究与实现-毕业论文（设计）.docVIP

. PAGE . 摘 要 为了解决交通问题，人们提出了智能交通系统(Intelligent Transportation System，简称ITS)。而智能交通系统与交通流量预测有密切的关系。在本课题中，首先分别用SVM方法，BP神经网络方法，Elman回归神经网络方法对快速路小时交通量进行预测。最后，采用组合预测的方法使用融合数据的技术综合对前面三种预测方法产生的预测值进行处理，选取适当的权值，最大程度的发挥各个模型的优点，分散单个预测模型的不稳定性，提高总体的确定性。此外，还获得了很多用单一模型预测不能得到的信息，最终提升预测准确度并得出最优预测结果，然后比较四种预测模型的预测结果的精度，以此来证明组合模型的优越性。 关键词 交通流量预测；支持向量机；BP神经网络；Elman神经网络；组合预测模型 Abstract Intelligent transportation system (ITS) has been proposed to solve the traffic problems. In this subject, SVM method, BP neural network method and Elman recurrent neural network method are used to predict the traffic volume per hour of the expressway at first. And combination forecasting method that use the technology of fusion data is adopted to process the generated prediction value generated by the three methods above Comprehensively at last, then, (we can) select appropriate weights to maximize the play of each model’s advantage in a large extent, disperse the instability of the single prediction model, improve the overall uncertainty. In addition, some information that can not be gotten by many single model forecasting is obtained, comparing the accuracy of the value that obtained by the four prediction models, then the superiority of the combination model is proved. Keywords: traffic flow forecasting; support vector machine; BP neural network; Elman neural network; combination forecasting model 目 录 TOC \o 1-3 \h \z \u HYPERLINK \l _Toc357984416 摘 要 PAGEREF _Toc357984416 \h 1 HYPERLINK \l _Toc357984417 Abstract PAGEREF _Toc357984417 \h 2 HYPERLINK \l _Toc357984419 第1章 绪论 PAGEREF _Toc357984419 \h 5 HYPERLINK \l _Toc357984420 1.1 选题背景 PAGEREF _Toc357984420 \h 5 HYPERLINK \l _Toc357984421 1.2 研究目的和意义 PAGEREF _Toc357984421 \h 5 HYPERLINK \l _Toc357984422 1.3 文献综述 PAGEREF _Toc357984422 \h 6 HYPERLINK \l _Toc357984423 1.4 论文的主要内容 PAGEREF _Toc357984423 \h 7 HYPERLINK \l _Toc357984424 第2章 基于支持向量机的快速路小时交通量预测 PAGEREF _Toc357984424 \h 8 HYPERLINK \l _Toc357984425 2.1 支持向量机的简介 PAGEREF _Toc