本文選題：神經(jīng)網(wǎng)絡(luò) + 遺傳算法　； 參考：《湖南大學(xué)》2012年碩士論文

【摘要】：作為判斷經(jīng)濟(jì)是否穩(wěn)定的先期指標(biāo),通貨膨脹率直接或間接影響著工資、利率、匯率等宏觀經(jīng)濟(jì)指標(biāo)。從上世紀(jì)90年代開始,將通貨膨脹率控制在合理水平,以促進(jìn)經(jīng)濟(jì)的平穩(wěn)發(fā)展逐漸成為各國(guó)中央銀行貨幣政策的中心目標(biāo)。在此背景下,通貨膨脹預(yù)測(cè)作為一種新的貨幣政策中介目標(biāo),在提高貨幣政策有效性,以穩(wěn)定物價(jià)方面發(fā)揮著越來(lái)越重要的作用,尤其是在實(shí)行通貨膨脹目標(biāo)制的國(guó)家。目前,雖然通貨膨脹預(yù)測(cè)方法眾多,但皆存在著不足之處,而且央行的決策也不僅僅參考一種模型結(jié)果,因此,構(gòu)建一種新的預(yù)測(cè)方法,為央行決策時(shí)提供對(duì)比參考便具有理論與現(xiàn)實(shí)意義。 本文遵循從理論闡述、實(shí)證分析到對(duì)策建議的思路展開。神經(jīng)網(wǎng)絡(luò)作為新興的預(yù)測(cè)方法,以其非線性、基于數(shù)據(jù)驅(qū)動(dòng)和穩(wěn)定性高等優(yōu)點(diǎn)在國(guó)外得到了廣泛應(yīng)用,國(guó)內(nèi)雖亦有眾多研究,但在通貨膨脹預(yù)測(cè)方面卻較少。因此本文在借鑒國(guó)內(nèi)外相關(guān)研究成果的基礎(chǔ)上,采用定性分析與定量分析相結(jié)合的方法,從理論和實(shí)證角度對(duì)神經(jīng)網(wǎng)絡(luò)在我國(guó)通貨膨脹預(yù)測(cè)方面的適用性進(jìn)行了研究。文章首先簡(jiǎn)要介紹了神經(jīng)網(wǎng)絡(luò)和通貨膨脹預(yù)測(cè)的基礎(chǔ)理論,對(duì)相關(guān)概念進(jìn)行了說明。然后羅列了在通貨膨脹預(yù)測(cè)方面經(jīng)驗(yàn)豐富的部分國(guó)家的工作流程以及特點(diǎn),對(duì)比分析了我國(guó)在此方面存在的不足。并利用我國(guó)通貨膨脹部分影響因素2005年3月到2011年12月的月度數(shù)據(jù)構(gòu)建了BP神經(jīng)網(wǎng)絡(luò)模型進(jìn)行短期預(yù)測(cè)。實(shí)證結(jié)果表明經(jīng)優(yōu)化后的網(wǎng)絡(luò)能根據(jù)已有數(shù)據(jù)對(duì)未來(lái)至少6個(gè)月的CPI進(jìn)行良好預(yù)測(cè),同時(shí)未經(jīng)優(yōu)化的網(wǎng)絡(luò)也給出了較為合理的預(yù)測(cè)結(jié)果。文章最后就如何提高央行通貨膨脹預(yù)測(cè)的準(zhǔn)確性提出了：提高統(tǒng)計(jì)數(shù)據(jù)質(zhì)量；增強(qiáng)中央銀行主動(dòng)搜集和分析信息的能力；完善我國(guó)的宏觀經(jīng)濟(jì)模型；疏通貨幣政策傳導(dǎo)渠道等相關(guān)政策建議。
[Abstract]:As an advance indicator to judge whether the economy is stable, the inflation rate directly or indirectly affects the macroeconomic indicators such as wages, interest rates, exchange rates and so on. Since the 1990s, keeping the inflation rate at a reasonable level to promote the smooth development of the economy has gradually become the central goal of the monetary policy of the central banks of various countries. In this context, inflation forecasting, as a new intermediate target of monetary policy, plays an increasingly important role in improving the effectiveness of monetary policy and stabilizing prices, especially in countries with inflation targeting system. At present, although there are many methods of predicting inflation, there are some shortcomings, and the decision of the central bank is not only based on the results of a model, therefore, a new forecasting method is constructed. It is of theoretical and practical significance to provide a comparative reference for central bank decision-making. This article follows from the theory elaboration, the demonstration analysis to the countermeasure suggestion train of thought unfolds. As a new forecasting method, neural network has been widely used in foreign countries because of its nonlinearity, data-driven and high stability. Although there are many researches in China, it is less in the field of inflation prediction. Therefore, based on the research results at home and abroad, this paper studies the applicability of neural network in China's inflation prediction from the perspective of theory and practice by combining qualitative analysis with quantitative analysis. In this paper, the basic theory of neural network and inflation prediction is introduced briefly, and the related concepts are explained. Then it lists the workflow and characteristics of some countries with rich experience in inflation forecasting, and compares and analyzes the shortcomings of our country in this respect. Based on the monthly data from March 2005 to December 2011, a BP neural network model is constructed to predict the inflation in China. The empirical results show that the optimized network can forecast CPI for at least 6 months according to the existing data, and the unoptimized network also gives a more reasonable forecast result. Finally, the paper puts forward how to improve the accuracy of the inflation forecast of the central bank: to improve the quality of statistical data, to strengthen the ability of the central bank to collect and analyze information actively, to perfect the macroeconomic model of our country; Dredging monetary policy transmission channels and other relevant policy recommendations.
【學(xué)位授予單位】：湖南大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2012
【分類號(hào)】：F822.5;TP18

【參考文獻(xiàn)】

相關(guān)期刊論文 前6條

1 胡志浩;;美聯(lián)儲(chǔ)如何制定貨幣政策?[J];銀行家;2008年02期

2 錢素萍;對(duì)統(tǒng)計(jì)數(shù)據(jù)質(zhì)量的幾點(diǎn)認(rèn)識(shí)[J];經(jīng)濟(jì)與管理;2004年11期

3 孫紅英;;2009年中國(guó)CPI運(yùn)行定量分析——基于改進(jìn)的BP神經(jīng)網(wǎng)絡(luò)[J];經(jīng)濟(jì)論壇;2009年20期

4 王宇;李旭東;李自力;;基于BP神經(jīng)網(wǎng)絡(luò)的我國(guó)CPI預(yù)測(cè)與對(duì)策[J];計(jì)算機(jī)科學(xué);2009年10期

5 薛永剛;;基于神經(jīng)網(wǎng)絡(luò)的通貨膨脹預(yù)測(cè)模型研究[J];商業(yè)時(shí)代;2010年18期

6 陳娟,余灼萍;我國(guó)居民消費(fèi)價(jià)格指數(shù)的短期預(yù)測(cè)[J];統(tǒng)計(jì)與決策;2005年04期

，

本文編號(hào)：2112190