【摘要】：隨著中國(guó)經(jīng)濟(jì)的高速發(fā)展和匯率市場(chǎng)化進(jìn)程的不斷加快,人民幣匯率在國(guó)內(nèi)外經(jīng)濟(jì)中的地位越來越重要。2015年8月11日,為增強(qiáng)人民幣匯率的市場(chǎng)化程度,讓人民幣順利加入特別提款權(quán),央行放棄了對(duì)每日人民幣兌美元中間價(jià)的干預(yù),匯率波動(dòng)幅度越來越大,短期匯率走勢(shì)也具有很強(qiáng)的不確定性。在此背景下,準(zhǔn)確預(yù)測(cè)匯率波動(dòng)性對(duì)降低中國(guó)企業(yè)的生產(chǎn)成本和規(guī)避匯率風(fēng)險(xiǎn)具有至關(guān)重要的理論研究意義和實(shí)際應(yīng)用價(jià)值。本文在國(guó)內(nèi)外學(xué)者對(duì)匯率的研究基礎(chǔ)上,根據(jù)匯率序列具有線性和非線性的復(fù)合特征,并利用ARIMA模型和GRNN模型分別在線性空間和非線性空間的預(yù)測(cè)優(yōu)勢(shì),并構(gòu)建了ARIMA與GRNN的組合模型對(duì)人民幣匯率波動(dòng)趨勢(shì)進(jìn)行實(shí)證分析。本文主要分為五部分:第一部分首先對(duì)人民幣匯率的研究背景及意義進(jìn)行了闡述,強(qiáng)調(diào)該課題研究的重要性,然后對(duì)人民幣匯率國(guó)內(nèi)外的研究動(dòng)態(tài)進(jìn)行了概述,主要包括國(guó)內(nèi)外學(xué)者對(duì)匯率波動(dòng)預(yù)測(cè)的時(shí)間序列分析法研究以及文獻(xiàn)評(píng)述,最后概括性地闡述了本文研究的主要內(nèi)容和方法。第二部分回顧了人民幣匯率制度的發(fā)展過程及不同制度下的人民幣兌美元匯率的變化趨勢(shì),并分析了人民幣兌美元匯率波動(dòng)給我們?nèi)粘Ｉ罴皣?guó)家經(jīng)濟(jì)帶來的影響。第三部分介紹了ARIMA模型和GRNN模型的相關(guān)理論以及模型參數(shù)設(shè)置的方法,緊接著闡述了ARIMA與GRNN的組合模型的原理及建模步驟。第四部分利用ARIMA與GRNN組合模型對(duì)人民幣匯率進(jìn)行實(shí)證分析。首先利用ARIMA模型對(duì)人民幣兌美元匯率中間序列進(jìn)行預(yù)測(cè)分析得到線性主體部分;然后利用GRNN神經(jīng)網(wǎng)絡(luò)模型對(duì)前一模型的殘差進(jìn)行預(yù)測(cè)分析得到非線性部分;最后將線性主體部分與非線性殘差部分相加得到人民幣兌美元匯率中間價(jià)序列的預(yù)測(cè)結(jié)果。研究結(jié)果表明,ARIMA-GRNN組合模型對(duì)人民幣兌美元匯率中間價(jià)的預(yù)測(cè)效果優(yōu)于單一的ARIMA模型和單一的GRNN模型,且GRNN模型比ARIMA模型的預(yù)測(cè)效果好。第五部分在對(duì)論文實(shí)證檢驗(yàn)的結(jié)果進(jìn)行分析總結(jié)的基礎(chǔ)上,還指出了論文的不足之處以及需要進(jìn)一步值得研究的地方。
[Abstract]:With the rapid development of Chinese economy and the accelerating process of marketization of exchange rate, RMB exchange rate is becoming more and more important in domestic and foreign economy. On August 11, 2015, to enhance the marketization of RMB exchange rate, Allowing the yuan to join the SDR smoothly, the central bank dropped its daily intervention against the dollar, the exchange rate fluctuated more and more, and short-term exchange rate movements were highly uncertain. Under this background, accurate prediction of exchange rate volatility is of great theoretical significance and practical application value in reducing production cost and avoiding exchange rate risk of Chinese enterprises. Based on the domestic and foreign scholars' research on exchange rate, according to the linear and nonlinear characteristics of exchange rate series, this paper makes use of the prediction advantages of ARIMA model and GRNN model in linear space and nonlinear space, respectively. The combination model of ARIMA and GRNN is constructed to analyze the fluctuation trend of RMB exchange rate. This paper is divided into five parts: the first part expounds the research background and significance of RMB exchange rate, emphasizes the importance of the research, and then summarizes the domestic and foreign research trends of RMB exchange rate. This paper mainly includes the research of time series analysis and literature review on the prediction of exchange rate fluctuation. Finally, the main contents and methods of this paper are summarized. The second part reviews the development process of RMB exchange rate system and the change trend of RMB / US dollar exchange rate under different systems, and analyzes the impact of RMB / US dollar exchange rate fluctuation on our daily life and national economy. The third part introduces the related theories of ARIMA model and GRNN model and the method of model parameter setting. Then the principle and modeling steps of the combined model of ARIMA and GRNN are expounded. The fourth part uses ARIMA and GRNN combination model to analyze RMB exchange rate empirically. Firstly, the linear principal part is obtained by using the ARIMA model to predict the intermediate sequence of RMB / US dollar exchange rate, and then the nonlinear part is obtained by using the GRNN neural network model to predict the residual of the former model. Finally, by adding the linear principal part and the nonlinear residual part, the prediction results of the intermediate price series of RMB / US dollar exchange rate are obtained. The results show that the ARIMA-GRNN combination model is superior to the single ARIMA model and the single GRNN model in predicting the intermediate price of RMB / US dollar exchange rate, and the GRNN model is better than the ARIMA model. On the basis of analyzing and summarizing the results of empirical test, the fifth part also points out the shortcomings of the paper and the points that need to be further studied.
【學(xué)位授予單位】：湖北工業(yè)大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：F832.6

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 張艷;蔡光興;;基于ARIMA和GRNN模型對(duì)人民幣匯率的預(yù)測(cè)[J];特區(qū)經(jīng)濟(jì);2017年02期

2 王正新;陳雁南;;“匯改”以來人民幣實(shí)際有效匯率波動(dòng)的非線性機(jī)制——基于平滑轉(zhuǎn)移自回歸模型的實(shí)證研究[J];金融理論與實(shí)踐;2016年06期

3 孔佳文;卞佳yN;方小萱;李文;;基于ARIMA模型的人民幣匯率分析及預(yù)測(cè)[J];現(xiàn)代經(jīng)濟(jì)信息;2016年11期

4 孫永利;王華金;郝麗;肖曉明;;基于神經(jīng)網(wǎng)絡(luò)和遺傳算法的螺旋折流板換熱器性能預(yù)測(cè)[J];化學(xué)工業(yè)與工程;2016年04期

5 宮舒文;;基于GARCH族模型的人民幣匯率波動(dòng)性分析[J];統(tǒng)計(jì)與決策;2015年12期

6 王建國(guó);;基于BP神經(jīng)網(wǎng)絡(luò)的股票價(jià)格反轉(zhuǎn)點(diǎn)預(yù)測(cè)[J];現(xiàn)代計(jì)算機(jī)(專業(yè)版);2015年05期

7 張瑩;邵毅;王式功;尚可政;李旭;劉慧;耿迪;;呼吸系統(tǒng)疾病死亡人數(shù)的人工神經(jīng)網(wǎng)絡(luò)方法研究[J];中國(guó)衛(wèi)生統(tǒng)計(jì);2014年05期

8 李明景;汪金菊;;基于ARMA-稀疏貝葉斯模型的匯率預(yù)測(cè)研究[J];合肥工業(yè)大學(xué)學(xué)報(bào)(自然科學(xué)版);2014年08期

9 危黎黎;李超;李余輝;;基于STAR模型的人民幣匯率非線性特征及預(yù)測(cè)[J];統(tǒng)計(jì)與決策;2014年09期

10 陳黎明;王春香;黃偉;胡晉武;;人民幣匯率波動(dòng)的非線性特征研究[J];統(tǒng)計(jì)與決策;2014年09期

相關(guān)博士學(xué)位論文 前1條

1 廖薇;基于神經(jīng)網(wǎng)絡(luò)和遺傳規(guī)劃的匯率預(yù)測(cè)技術(shù)研究[D];華東師范大學(xué);2010年

相關(guān)碩士學(xué)位論文 前2條

1 陳天舒;基于ARIMA與GPR組合模型的人民幣匯率預(yù)測(cè)[D];山東大學(xué);2015年

2 黃臘梅;人民幣匯率GARCH-GRNN組合預(yù)測(cè)研究[D];湖南大學(xué);2009年

，

本文編號(hào)：2253554