基于因子分析和Logistic模型的中国客运量预测

China's Passenger Traffic Forecast based on Factor Analysis and Logistic Model

通过SPSS软件对1994-2011年中国客运量及影响因素进行因子分析，建立了中国客运量为被解释变量，其他影响因素为解释变量的多元统计分析模型。通过因子分析法分析影响中国客运量的主要因素，并进行了降维处理，消除由于变量过多而导致的多重共线性影响，同时得到反映不同年份的综合经济发展值的因子得分。考虑到2003年中国由于“非典”而导致的客运量降低的异常值的出现，将综合经济发展值与时间的关系首先作为一个整体研究，然后分为2个时间段分别研究，分别求出不同的综合经济发展值，构建了基于Logistic模型的中国客运量预测模型，进一步预测未来中国客运量。预测结果表明，通过较长时间段得到的综合经济发展值预测的数据误差相对较小。

This paper analyzes China＇s passenger traffic from 1994 to 2011 and its impact factors through SPSS by building a multivariate statistical analysis model in which China＇s passenger traffic serves as the dependent variable and the factors as independent variables. The main factors influencing the passenger traffic in China are analyzed in detail based on the factor analysis and are rearranged through dimension reduction approach with the purpose of eliminating the multi collinearity resulting from variables overloading. Considering the special event of ＂ SARS＂ in 2003 which resulted in abnormal value of passenger volume, we study the relationship between comprehensive economic development and time series as a whole, and conduct individual research by dividing the time periods into two. Furthermore, a logisticbased forecast model is set up accordingly to further predict the future passenger traffic in China. The empirical results show that the longer-period comprehensive economic development value has less forecasting error.