宏观经济视角下基于多维灰色组合模型的汇率预测研究  

Research on Exchange Rate Prediction Based on Multidimensional Gray Combination Model from Macroeconomic Perspective

作　　者：姚天祥 刘熙纯 YAO Tianxiang;LIU Xichun

机构地区：[1]南京信息工程大学管理工程学院

出　　处：《阅江学刊》2023年第4期83-92,172,共11页Yuejiang Academic Journal

基　　金：江苏省生产力学会资助项目“‘双循环’格局下的中国经济增长预测研究”(JSSCL2020A011)。

摘　　要：随着人民币汇率市场化进程的加快,汇率波动的弹性空间也越来越大。准确把握人民币汇率走势对于我国经济社会稳健发展具有重要意义。在宏观经济视角下,对2010年7月至2022年8月的人民币兑换美元月度汇率中间价进行了数据拟合,发现由灰色系统模型和支持向量回归模型并联组合得到的多维灰色组合模型具有更好的拟合效果,进而利用该模型预测2022年9月至2023年5月人民币兑换美元汇率月度价格。预测结果显示,2022年9月至2022年12月,受到外资流出和美元加息影响,人民币兑换美元汇率仍存在持续贬值压力,但在2023年初将出现拐点,汇率企稳回升,该结果与人民币兑换美元汇率的实际走势较为吻合,进一步验证了多维灰色组合模型在汇率预测方面具有优势。结合国内外经济发展形势来看,人民币币值短期贬值并不会影响其长期稳定性。As the marketization of RMB exchange rate accelerates,the elasticity of exchange rate fluctuation is becoming larger and larger.It is important to accurately grasp the trend of RMB exchange rate for the sound development of China’s economy and society.In the paper,we fit data on the monthly median exchange rate of RMB to USD from July 2010 to August 2022 under the macroeconomic perspective,and find that the multidimensional gray combination model obtained by the parallel combination of gray system model and support vector regression model has better fitting effect,and then use the model to forecast the monthly price of RMB to USD exchange rate from September 2022 to May 2023.The prediction results show that from September 2022 to December 2022,there will still be continuous depreciation pressure on the RMB to USD exchange rate due to foreign capital outflow and USD interest rate hike,but there will be an inflection point at the beginning of 2023,and the exchange rate will stabilize and rebound,which is highly consistent with the actual trend of RMB to USD exchange rate,further verifying the advantage of the multidimensional gray combination model in exchange rate prediction.Combined with the domestic and international economic development situation,the short-term depreciation of the RMB currency does not affect its long-term stability.

关 键 词：多元融合模型 宏观经济分析 灰色系统 支持向量回归 汇率预测 

分 类 号：N941.5[自然科学总论—系统科学] F832.6[经济管理—金融学]