论文部分内容阅读
根据影响城市轨道交通节假日一票通进站量因素的特点,把自变量分为连续变量和虚拟变量。依据虚拟变量的有序性和无序性,实现不同的量化方法,建立轨道交通节假日一票通进站量与连续变量和虚拟变量的回归方程。对回归方程的残差进行平稳性检验,建立残存序列的ARMA(1,5)模型。结合以上两个模型,构建基于虚拟变量的回归方程和ARMA(1,5)的组合预测模型。以北京2014年10月1日到2016年4月4日之间所有节假日一票通进站量数据进行实例分析,验证了组合模型的精度比原来的回归模型的精度提高了12.33%,说明此组合预测模型是有效的。
According to the characteristics of the one-pass and one-stop factors affecting the urban rail transit holiday, the independent variables are divided into continuous variables and dummy variables. According to the orderliness and disorder of the dummy variables, different quantification methods are realized, and the regression equation of one-ticket pass-in stations and the continuous variables and dummy variables of the rail transit holiday is established. The residuals of the regression equation are tested for stationarity, and the ARMA (1,5) model of the residual sequence is established. Combining the above two models, a combined forecasting model based on virtual variables and ARMA (1,5) is constructed. An example analysis was conducted on the data of one-way pass from Beijing October 1, 2014 to April 4, 2016, which verified that the accuracy of the combined model was 12.33% higher than the original regression model, indicating that the combined forecast The model is valid.