基于动态组合模型的我国私人汽车拥有量预测分析

发布时间：2018-03-01 20:37

本文关键词： 私人汽车拥有量时间序列偏最小二乘动态组合　出处：《四川师范大学》2017年硕士论文　论文类型：学位论文

【摘要】：1985年,我国私人汽车拥有量仅有28.49万辆,经过三十年的发展,到2015年底,我国私人汽车拥有量已超过14000万辆。在可预见的未来,私人汽车拥有量也将保持高速、持续增长。但是,私人汽车在方便人们出行的时候,也带来了一系列亟待解决的社会问题,比如日益严重的交通拥堵、空气污染等问题。因此,预测我国私人汽车拥有量及其变化趋势,已经成为研究的热点。科学准确的预测私人汽车拥有量,能为国家决策部门做出相应的产业政策调整、供给侧改革提供理论基础和数据支撑,同时也能够为汽车行业的规划发展方向提供依据。本文围绕我国私人汽车拥有量及其主要影响因素展开研究。研究的目的是通过建立动态组合模型来预测我国私人汽车的拥有量及其发展趋势,并以此为出发点,提出合理的建议。为此,主要做了以下研究工作:首先,根据偏最小二乘回归理论,建立私人汽车拥有量与其主要因素之间的回归模型。同时,通过对各影响因素进行预测,进而利用偏最小二乘回归模型对私人汽车拥有量做出了预测。其次,通过对私人汽车拥有量的时序数据的分析,建立时间序列季节模型,并利用季节模型预测私人汽车的拥有量。最后,以回归组合预测法来确定组合权重,同时,考虑到单项预测模型可能会随着时间的变化,预测效果将会发生改变,因此,加入了时间参数,对单项预测模型进行修正,采用可变加权系数,建立动态组合预测模型,从而对我国私人汽车未来发展做出更合理的预测。本文的贡献在于:一方面,在分析私人汽车拥有量与其影响因素之间的关系时,采用了偏最小二乘回归方法。该方法在提取成分,在有效地解决影响因素之间多重共线性问题的同时,又最大限度地使提取的成分与私人汽车拥有量有关。另一面,考虑到单项预测模型可能会随着时间的变化,预测效果将会发生改变,因此,增加了时间参数对单项预测模型进行修正,从而建立的动态组合预测模型,在实际研究中更具有研究和运用价值。
[Abstract]:In 1985, there were only 284,900 private cars in our country. After 30 years of development, by end of 2015, the number of private cars in our country had exceeded 140 million. In the foreseeable future, private car ownership will also remain at a high speed. However, when private cars are convenient for people to travel, they also bring about a series of social problems that need to be solved urgently, such as increasing traffic congestion, air pollution and so on. Forecasting private car ownership and its changing trend in China has become a hot topic of research. Scientific and accurate prediction of private car ownership can make corresponding industrial policy adjustments for national policy-making departments. The supply-side reform provides the theoretical basis and data support, At the same time, it can also provide the basis for the planning and development direction of automobile industry. The purpose of this study is to establish a dynamic combination model to predict China's private car ownership and its main influencing factors. The ownership of people and cars and its development trend, Based on the partial least square regression theory, the regression model between the ownership of private cars and its main factors is established. At the same time, the following research work is done: first, according to the partial least squares regression theory, a regression model between the ownership of private cars and its main factors is established. Through the prediction of the influencing factors, the partial least square regression model is used to predict the ownership of private cars. Secondly, through the analysis of the time series data of the ownership of private cars, the seasonal model of time series is established. Finally, the regression combination forecasting method is used to determine the combination weight, and considering that the single prediction model may change over time, the forecasting effect will change. The time parameter is added, the single prediction model is modified, the variable weighting coefficient is adopted, and the dynamic combination forecasting model is established, which makes a more reasonable prediction for the future development of private cars in China. The contributions of this paper are as follows: on the one hand, The partial least square regression method is used to analyze the relationship between private car ownership and its influencing factors. On the other hand, considering that the single prediction model may change over time, the prediction effect will change. The time parameter is added to modify the single prediction model, and the dynamic combination prediction model is established, which has more research and application value in the actual research.
【学位授予单位】：四川师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O212.1

【参考文献】