【作者】 雷晨；

【导师】 余燕春；

【作者基本信息】 浙江大学 ， 国际贸易学， 2010， 硕士

【摘要】 二十一世纪以后,国际原油价格逐渐呈现出不同以往的运动规律,其价格短期波动趋势明显加剧,这一现象已经难以运用传统的长期价格影响因素进行合理有效的解释。在这一背景下,研究国际原油价格的短期波动规律具有重要的意义。本文通过运用Hodrick-Prescott滤波方法对原油价格波动中短期因子进行分离,以此为基础运用计量经济方法对外部影响因素和原油价格短期波动的相互作用关系进行研究,并构建了基于径向基函数神经网络的原油价格短期波动预测模型。与国际原油价格日趋动荡相对应的是,由于我国油品价格机制的改革以及原油对外依存度的提高,共同导致我国经济发展越来越多受到国际原油市场的影响,因此本文进一步采用Granger因果检验等计量研究方法对国际原油价格短期波动与我国宏观经济的相互作用关系,以及原油价格对我国产生影响的传导路径进行了深入分析。本研究主要得到以下研究结论：(1)国际原油价格短期波动具有强记忆性(国际原油价格对自身波动解释的贡献率达到80%),此外美元汇率也是解释原油价格短期波动的重要变量。(2)基于径向基函数神经网络预测模型其拟合效果显著优于传统的多元线性回归预测模型,从侧面证实了外部作用因素对国际原油价格短期波动的影响具有非线性特点。(3)国际原油价格短期波动对我国居民消费价格指数和银行间实际回购利率变化具有直接的解释作用,但并未发现实际工业增加值增长和广义货币供应量增长会受到其任何直接的影响。更多还原

【Abstract】 After the twenty-first century, the international crude oil prices have rendered a different movement style, of which Short-term fluctuations is clearly increasing. It is hard to employ the traditional long-term factors to explain this phenomenon reasonably and effectively. Based on this special background, this research applied Hodrick-Prescott filter method to remove the trends from time series data. And then tried to study the interaction between the influencing factors and the short-term fluctuations in crude oil prices using Econometric Methods. Finally, we constructed short-term fluctuations in crude oil price forecasting model based on radial basis function neural network.With the increasingly volatile international crude oil prices, the reform of China’s oil pricing mechanism, as well as the increased dependence on foreign crude oil result in that China’s economic development is affected by the international crude oil market more and more. Therefore, this article further use Granger causality test and other research methods to examine the relation between short-term fluctuations in international crude oil prices and China’s macro-economy and discuss the transmission path, relevant research conclusions are as follows:(1) The short-term fluctuations in international crude oil prices have a strong memory, in addition, the exchange of U.S. dollar is also an important Explanatory variables.(2) Fitting result of the forecasting model based on radial basis function neural network is significantly better than the traditional multiple linear regression prediction model, which confirms that external factors on the fluctuations of international crude oil prices in short-term have non-linear characteristics.(3) Short-term fluctuations in international crude oil price have a direct interpretation of the role of China’s consumer price index and the real inter-bank repo interest rate, however we didn’t find Short-term fluctuations shows any direct impact on real industrial value added and broad money supply.更多还原