摘要
进出口贸易在我国国民经济中所占的地位越来越重要,它是否能健康平稳地发展直接关系到国家的竞争实力。根据预测的结果,可以正确评估国内经济发展状况,制定适合我国国情的进出口政策和对外贸易政策,有利于顺利发展我国经济。但由于我国外贸内部环境和外部环境的不确定性因素较多,导致了外贸系统的复杂性;加之外贸历史数据不齐全,无疑在预测实践中加大了可操作性的难度;同样也造成了选取何种预测方法的困扰。因此,选取适合我国国情的预测方法,从而可以较精确地预测出进出口贸易总额,具有重大意义。
本文首先介绍了外贸预测的一般思路,分别是确定预测目标、收集预测资料、选择预测方法和评价预测结果。所有具体的预测方法在实际应用在思路上都应如此,只是在原理和操作上有所区别。
接着,关于所有对我国外贸进出口额进行预测的预测方法,本文一一作了介绍。所列举的预测方法,大体上可分为定性预测法和定量预测法;具体的方法有:定性预测法、逐步回归预测法、计量经济模型法、自回归移动平均法、人工神经网络法、灰色建模预测法六种。每种方法基于不同的数学理论,使用的模型也不尽相同。因此,它们在对数据的要求、预测精度和预测期间上也有差异。
然后,文章介绍了每种预测方法的理论基础和数学模型,以及相应的操作步骤;并将其应用到对我国某段时间的外贸预测中去,利用历史数据,得出未来某个时段内贸易额的数据。
在之后进行的对各预测方法的比较分析中,先针对各预测原理的特点进行相应的比较,区分各个预测方法的优势与劣势;然后对预测结果的预测期间和预测精度进行比较。预测精度又与预测时间长短有密切关系。一般而言,预测时间越短,影响预测结果的因素的变化越小,预测误差也越小。反之,预测时间越长,影响预测结果因素的变化也越大,产生的误差也越大。
在总结中,通过比较各预测方法的理论基础、各个模型的优劣,以及预测精度和预测的时间跨度,并结合数据的可获得性,从而得出:适合我国外贸的预测方法,要从预测目的以及期望的预测精度出发,根据各个模型的特点,并结合历史数据的可获得性,来选取适当的预测方法。
The import and export trade has been playing an increasingly important part in our national economy. Whether it can smoothly develop has been immediately concerned with China's competitive forces. According to the predicting resultes, we can make correct evaluation about domestic economy and, set up foreign trade policy which suites to China's actual conditions, so that it's favourable to the development of our foreign economy and trade. But there are many uncertain elements in China's internal and external environments, which would bring complexity to the foreign trade system. Also, the incomplete historical data of our foreign trade would add difficulty to the predicting practice, as well as the uncertainty to what predicting method should be chosen. Therefore, selecting a suitable method for national situations, making a precise prediction of total imports and exports would make great significance.
This paper firstly has introduced a general idea to predict foreign trade, which contains setting predicting purposes; collecting predicting informations; choosing predicting method and evaluating predicting results. All the specific predicting practice should follow this way in applications, except in principles and operations.
Next, this paper makes a reference to all the methods predicting China's imports and exports. The methods generally include quantitative and qualitative predicting. Concretely , there are quantitative predicting ; Multi-hierarchic Recursive Regression Algorithm; econometric models; Autoregressive Moving Average Model(ARIMA); Artificial Neural Network and GM(1,1). Each method is based on different mathematical theories or models. So everyone has differences in the requirement of data, also has defferences in the predicting precision and period.
Then, the theories and the mathematical models of every predicting methods have been introduced in this paper, as well as their corresponding operational procedures. After that, methods would be used to predict China's foreign trade volume. The historical data would also be used to gain the trade volume during a certain period in the future.
Followed up with the analysing and comparing of the predicting methods, the characteristic of every predicting theory will be discussed firstly. Then the advantages and disadvantages of each predicting method.will be distinguished. After that, the predicting precision and the predicting period will also be compared. Generally speaking, the shorter the predicting period is, the greater changes and errors would be reacted in the results.
In the summary, by discussing theories; characteristics; precision; period of the predicting methods involved above, we can draw a conclution that: in order to choose a favorable predicting method, it's better to proceed from the predicting purposes and precision and, according to the characteristics of each model, relating to the availability of historical data, and then select a proper predicting method.
引文
[1] 赵伟.中国外贸:2002年扩张的环境与前景.商业经济与管理,2002,123(1):5~10
[2] 秦殿军,郭晓晶.对中国外贸形势的分析与预测.北方经贸,2002(7):74~75
[3] 钟俊.2005年度中国对外贸易预测与形势分析.南方经济,2005(1):72~74
[4] 国家经济信息中心经济预测部.经济论坛,2005(3):33~34
[5] 王红,童恒庆和魏平.进出口贸易额预测的逐步回归建模研究.统计与决策,2006(1):80~82
[6] 吴为英,向阳,崔家保.改进的多层递阶回归算法在进出口贸易中的应用.山东科技大学学报(自然科学版),2003,22(1):54~57
[7] 王振全,徐山鹰.协整模型与预测——中国外贸出口的定量研究.国际贸易,2000(1):23~26
[8] 张一,徐山鹰,汪寿阳.一类基于神经元网络的误差纠正模型的应用——2003年度中国出口预测.预测,2003,22(3):21~26
[9] 傅晓旗,谢雯,郑桂环.2006年中国进出口预测与分析.管理评论,2006,18(1):22~26
[10] George E.p.Box,Gwilym M.Jenkins,Gregory C.Reinsel.时间序列分析:预测与控制(英文版·第3版).人民邮电出版社,2005
[11] 郝雁.中国的进出口与经济增长:ARMA最优预测模型分析.统计与决策,2005(11):78-80
[12] 石柱鲜,王威,王立勇.对我国2005-2006年经济增长态势的分析与预测.经济与管理研究,2005(11):7~10
[13] 张家颖,龚玉策.基于时间序列方法我国2005年进出口总额预测分析.广东白云学院学刊,2005(2):59~64
[14] 王玉荣.ARIMA模型在我国出口贸易预测中的应用.决策参考,2004(4):33~34
[15] 祝树金,赖明勇.基丁贝叶斯正则化的TDBPNN模型在中国外贸预报中的应用及评估.中国管理科学,2005,13(1):1~8
[16] 邓聚龙.灰色预测与决策.武汉华中理工大学出版社,1986
[17] 王春芝,周文军.进山口贸易预测的灰色建模方法研究.工业技术经济,2002(6):76-77
[18] 邹晶,姜志新.基于GM(1.1)的灰色系统模型在我国外贸出口预测中的应用.国际贸易问题.2004(2):27~31
[19] Engle, R. E, C. W. J. Granger. Cointegration and error correction: representation, estimation and testing. Econometrica, 1987, 55(2):251~276
[20] 刘斌,邓述慧.中国的货币需求函数的非线性建模与预测.系统工程理论与实践,1997,17(4):40~57
[21] 魏巍贤.民币汇率的稳定机制及其动态过程.系统工程理论与实践,1999,19(8):54~61
[22] 马海军,贾湖,盛昭瀚.非线性混沌经济时序的预测方法及其应用研究.管理科学学报,2001(4):12~16
[23] 陈平.文明分岔经济混沌和演化经济学.经济科学出版社,2000
[24] 胡玉霞,高金峰.一种预测混沌时间序列的模糊神经网络方法.物理学报,2005(11):87~92
[25] 张一,陈振峰,徐山鹰.2004年度中国对外贸易预测与形势分析.预测,2003,15(12):3~12
[26] 刘华杰,王军.经济系统中的非线性与混沌及其应用.数量经济技术经济研究,1993(2):18~22
[27] 黄梦桥,王涛生.国际贸易系统中的分数维和最大Lyapunov指数.经济数学,2005,22(3):301~306
[28] 李树.混沌理论在经济系统中的应用.技术经济与管理研究,1998(5):45~48
[29] Meese, Ro se. An Empirical Assessment of Nonlinearities in Modes of Exchange Rate Determination. Re view of Economic Studies, 1991(5)
[30] Grassberger P. and I. Procaccia. Measuring the Strange needs of Strange Attractors. Physica D, 1983(9)
[31] Stutzer, Michael. Chaotic Dynamics and Bifurcation in a MacroModel. Journal of Economic Dynamics and Control, 1980(2)
[32] Day R. Irregular Growth Cycles. Am. Econ. Rev, 1982(72)
[33] Henry D I, ReggieB, JamesB K. Predict ion in chaotic nonlinear system: methods for time series with broadband Fourier spectra. Phys. Rev A, 1990(4)
[34] Rosser J B. Long Wave Chaos and Systematic Economic Transformation. World Future, 1994(99)
[35] Baumol WJ, Benhabib J. Chaos: Significance, mechanism and economic application, 1989(1)
[36] Levy D. Chaos Theory and Strategy: Theory, Application and Managerial Implications, 1994(15)
[37] Foresee F. Dan and Hagan M. J. Gauss-Newton approximation to Bayesian Learning. Proceedings of the international Joint Conference on Neural Networks, 1997(6)
[38] Wolf, A., J. B. Swift, H. L. Swinney, J. A. Vastano. Determining Lyapunov exponents from a time series. Physica D, 1985(16)
[39] Lapedes R. and R. Farber. Nonlinear Signal Processing Using Neural Networks: Prediction and System Modeling. Los Alamos National Laboratory, 1987
[40] MacKay, D. J. C. Bayesian. Interpolation. NeuralComputation, 1992(4)
[41] Harold Hotelling. Analysis of a complex of statistical variables into principal components. Journal of Educational Psychology, 1933(24): 212~315
[42] Porter, Michael E. Competitive Strategy: Techniques for Analyzing Industries and Competiotor, Free Press, NewYork, 1980
[43] GaryM G, Nabeel, M ark S. Recurrence matrices and the preservation of dynamical properties. Phys. Lett A, 1997(12)