论文部分内容阅读
在重庆市政府回购、租赁高速公路的背景下,研究了如何合理、准确地预测高速公路交通量,从而为政府部门及高速公路投资者的投资决策提供依据。根据高速公路年交通量样本小、预测期长、受经济因素影响等特点,选用了支持向量机回归来进行多因素单目标的预测。在预测过程中,为了提高精度,首先将所搜集的经济因素进行主成分分析,对指标进行了约减;然后用PSO方法对支持向量机参数进行了优化。
Under the background of repurchase and leased expressway in Chongqing municipal government, this paper studies how to predict the expressway traffic volume rationally and accurately so as to provide the basis for the investment decision of government departments and highway investors. According to the characteristics of freeway annual traffic samples, long forecasting period and economic factors, SVM regression is used to forecast the multi-factor single-objective. In the prediction process, in order to improve the accuracy, the principal components of the collected economic factors are firstly analyzed, and the indexes are reduced by a factor of approximately. Then, the PSO method is used to optimize the parameters of the SVM.