【作者】 王菲菲；

【导师】 马超群；

【作者基本信息】 湖南大学 ， 技术经济及管理， 2007， 硕士

【摘要】 目前,预测已经成为一门重要的学科。伴随着预测科学的发展,非线性预测方法的使用及组合预测方法是目前研究的重点,同时也随之出现了很多的非线性方法和组合预测方法,但是,不同的事物因所受影响因素的不同,表现出不同的发展规律,而不同的组合预测方法因其构成机理的不同而导致适用条件的不同。预测模型的建立其根本目的就是为了解决不同的预测问题,因此对不同的组合预测模型进行分析,发现现有组合预测模型的不足之处;根据事物发展规律选择合适的组合预测模型以提高预测精度,具有非常重要的作用和意义。本文旨在构建合适的组合预测模型对我国快速发展的区域出口贸易额进行预测,为未来的区域出口贸易发展和调控提供依据。本文在对组合预测、及出口预测的相关文献进行分析的基础上,提出了现有研究存在的主要问题,接着本文对区域出口贸易额预测模型的构建机理进行分析,在结合组合预测模型的优缺点以及区域出口贸易额变化特点的基础上,选择混合预测模型为区域出口贸易额的预测模型。然后构建了由误差校正-向量自回归模型以及支持向量机方法相混合的区域出口贸易额的组合预测模型。本文的最后一章对所建立的区域出口贸易额的混合预测模型进行了实证研究,以湖南省的出口贸易额为例,选取1999.01-2006.06月份的出口贸易额,首先运用误差校正-向量自回归模型,选取工业总产值、财政支出、社会消费品零售总额以及相关的汇率指标、广东省的出口贸易额等指标等建立了湖南省区域出口贸易额的线性预测模型,并运用SVM方法对模型中的协整向量进行非线性化,从而建立湖南省出口贸易额的非线性组合预测模型,并对2006.07-2007.06期间的湖南省出口贸易额进行预测。结果表明,文中所建立的组合预测模型的预测精度明显优于线性VEC模型、BP方法、SVM方法。最后对未来一年的湖南省月度出口贸易额进行了预测。更多还原

【Abstract】 At present, the forecast has become an important subject. Along with the development of forecasting theory, it is popular to use nonlinear forecasting methods and combined forecasting method. But different things were affected by different factors, showing different development, and different combined forecasting model has different application conditions because of different construction mechanism. The fundamental purpose of forecast is to solve different problems, so it is very important role and significance to analysis different combined forecasting model, found disadvantages, choose suitable combined forecasting mode according to the development of things improve prediction accuracy. This paper seeks to make a suitable combined forecasting model on Chinese regional export trade and forecast, to provide a reference for regulation.In this paper, the author first analysis the relevant literature about combined forecasting and export forecasting,in this basic ,the paper proposed problems. Secondly, this paper analysis the model’s construction principle of regional export trade, considered the features of kinds of combined forecast methods and characteristics of export trade, and chose hybrid forecast model to be the forecasting model of regional export. Then this paper constructed a hybrid forecasting model of region export trade based on VEC model and SVM method.The last chapter of this paper is empirical study. This chapter established a hybrid forecasting model of regional export trade in Hunan Province as an example, selected 1999.01-2006.06 regional export trade as an example, the first use VEC model, selected factors of the total industrial output, financial expenditure, the total retail sales of social consumer goods and the relevant exchange rate targets, the export trade of Guangdong Province, and other indicators, used the hybrid forecasting model that we constructed above, to give an non-linear hybrid forecasting model of Hunan Province’s export trade, and forecasted export trade of on 2006.07-2007.06. The resulted showed that this non-linear hybrid forecast model is better than linear model, BP method and SVM method. At last, the paper forecast Hunan export trade’s month volume in next year.更多还原