本文選題：主成分回歸 + 灰色關(guān)聯(lián)分析　； 參考：《華中師范大學(xué)》2017年碩士論文

【摘要】：在當(dāng)今世界,國(guó)家的發(fā)展離不開經(jīng)濟(jì)的發(fā)展,擴(kuò)大消費(fèi)需求是促進(jìn)經(jīng)濟(jì)發(fā)展的重中之重,而社會(huì)消費(fèi)品零售總額是反映一個(gè)國(guó)家人們消費(fèi)水平的重要因素,在國(guó)民經(jīng)濟(jì)體系中,也是一個(gè)很重要的指標(biāo)。因此,分析研究社會(huì)消費(fèi)品零售總額對(duì)于處在轉(zhuǎn)型期的中國(guó)經(jīng)濟(jì)更好的發(fā)展具有重要的意義。首先,本文以1991年至2015年近25年的時(shí)間序列數(shù)據(jù)為依據(jù),以Eviews 8.0為工具,對(duì)取對(duì)數(shù)后的數(shù)據(jù)進(jìn)行了主成分分析,根據(jù)累計(jì)貢獻(xiàn)率,選取出幾乎反映原變量所有信息的第一主成分F1t,再對(duì)選取出的主成分和社會(huì)消費(fèi)品零售總額做主成分回歸模型,最終得出主成分回歸結(jié)果為:Yt=0.8308F1t+1.2145AR(1)-0.4614AR(2)。緊接著又用近 10 年的時(shí)間序列數(shù)據(jù),運(yùn)用 Matlab 編程進(jìn)行了灰色關(guān)聯(lián)度分析,得到如下結(jié)論:我國(guó)社會(huì)消費(fèi)品零售總額與居民可支配收入、國(guó)內(nèi)生產(chǎn)總值、居民消費(fèi)水平有很大關(guān)系,而與人口數(shù)量、物價(jià)指數(shù)、恩格爾系數(shù)關(guān)系不大,最后在灰色關(guān)聯(lián)度分析的基礎(chǔ)上建立了灰色GM(1,5)模型,得出主要影響因素對(duì)社會(huì)消費(fèi)品零售總額的影響水平。然后,本文分別以社會(huì)消費(fèi)品零售總額1952年至2016年的年度數(shù)據(jù)以及2001年1月至2016年12月的月度數(shù)據(jù)為依據(jù),根據(jù)自相關(guān)和偏自相關(guān)函數(shù)圖進(jìn)行模型識(shí)別,運(yùn)用Eviews 8.0建立ARMA模型,并根據(jù)各種檢驗(yàn)最終確定最優(yōu)ARMA模型,最終在對(duì)社會(huì)消費(fèi)品零售總額的年度數(shù)據(jù)預(yù)測(cè)中,建立了ARIMA(1,1,0)模型,并通過(guò)2011年至2016年六年的實(shí)際值,預(yù)測(cè)值之間的誤差說(shuō)明預(yù)測(cè)結(jié)果的理想性,接著再利用此模型對(duì)2017年至2020年未來(lái)4年的社會(huì)消費(fèi)品零售總額進(jìn)行預(yù)測(cè)。在對(duì)社會(huì)消費(fèi)品零售總額的月度數(shù)據(jù)預(yù)測(cè)中,本文最終建立的是ARIMA(2,2,1)(0,1,1)12模型,同樣通過(guò)各種檢驗(yàn)說(shuō)明所建模型為最優(yōu)模型,并運(yùn)用此模型對(duì)2017年的各月度社會(huì)消費(fèi)品零售總額進(jìn)行預(yù)測(cè),進(jìn)而再對(duì)預(yù)測(cè)結(jié)果進(jìn)行分析。本文的研究結(jié)果對(duì)當(dāng)今經(jīng)濟(jì)的發(fā)展具有一定的實(shí)際意義,有利于制定宏觀經(jīng)濟(jì)政策,促進(jìn)國(guó)家發(fā)展。
[Abstract]:In today's world, the development of the country is inseparable from the development of the economy. Expanding consumption demand is the most important thing in promoting economic development, and the total amount of retail sales of social consumer goods is an important factor reflecting the consumption level of people in a country, and in the national economic system, Is also a very important indicator. Therefore, it is of great significance to analyze and study the total retail sales of consumer goods for the better development of Chinese economy in the transition period. Firstly, based on the time series data of nearly 25 years from 1991 to 2015 and using Eviews 8.0 as a tool, this paper analyzes the logarithmic data by principal component analysis, according to the cumulative contribution rate. The first principal component F1t, which almost reflects all the information of the original variable, is selected, and then the principal component regression model is made for the selected principal component and the total retail sales of consumer goods. Finally, the result of principal component regression is: Ytt 0.8308F1t 1.2145AR (1) -0.4614AR (2). Then, with the time series data of nearly 10 years, the grey correlation degree analysis is carried out with Matlab programming, and the following conclusions are obtained: the total retail sales of consumer goods in China and the disposable income of the residents, the gross domestic product (GDP), There is a great relationship between the consumption level of residents, but not with the population quantity, price index, Engel coefficient. Finally, the grey GM (1 / 5) model is established on the basis of grey correlation degree analysis. The influence level of the main influencing factors on the total retail sales of consumer goods is obtained. Then, based on the annual data of total retail sales of consumer goods from 1952 to 2016 and the monthly data from January 2001 to December 2016, according to the autocorrelation and partial autocorrelation function diagrams, the ARMA model is established by using Eviews 8.0. According to various tests, the optimal ARMA model is determined. Finally, in the annual data forecast of the total retail sales of consumer goods, Arima model is established, and the actual value of six years from 2011 to 2016 is obtained. The error between the predicted values shows the ideal of the predicted results, and then uses this model to predict the total retail sales of consumer goods in the next four years from 2017 to 2020. In the monthly forecast of the total retail sales of consumer goods, the Arima (2 / 2 / 1) (0 / 1 / 1) 12 model is established in this paper. The model is also proved to be the optimal model through various tests. The model is used to forecast the monthly retail sales of consumer goods in 2017, and then the forecast results are analyzed. The results of this paper are of practical significance to the economic development of today, which is conducive to the formulation of macroeconomic policies and the promotion of national development.
【學(xué)位授予單位】：華中師范大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：F724

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 張洪軍;;主成分分析方法求解主成分方法的改進(jìn)[J];學(xué)術(shù)問(wèn)題研究;2016年01期

2 龍會(huì)典;嚴(yán)廣樂(lè);;基于SARIMA、GM(1,1)和BP神經(jīng)網(wǎng)絡(luò)集成模型的GDP時(shí)間序列預(yù)測(cè)研究[J];數(shù)理統(tǒng)計(jì)與管理;2013年05期

3 張先兵;張松;;我國(guó)人均國(guó)內(nèi)生產(chǎn)總值的時(shí)間序列分析(1952-2011)與靜態(tài)預(yù)測(cè)(2012-2015)[J];現(xiàn)代管理科學(xué);2013年06期

4 首招勇;楊媛媛;;時(shí)間序列問(wèn)題的建模方法和過(guò)程[J];數(shù)學(xué)理論與應(yīng)用;2012年01期

5 肖元真;汪寶平;劉婷婷;;我國(guó)應(yīng)對(duì)全球金融危機(jī)的戰(zhàn)略抉擇[J];中國(guó)國(guó)情國(guó)力;2009年03期

6 劉亞男;;社會(huì)消費(fèi)品零售總額的影響因素分析[J];時(shí)代經(jīng)貿(mào)(中旬刊);2007年SB期

7 張曉峰;李博;;ARIMA模型在社會(huì)消費(fèi)品零售總額預(yù)測(cè)中的應(yīng)用[J];商場(chǎng)現(xiàn)代化;2007年32期

8 王耀青;;對(duì)我國(guó)社會(huì)消費(fèi)品零售總額的分析和預(yù)測(cè)[J];太原科技大學(xué)學(xué)報(bào);2007年05期

9 李巧梅;熊國(guó)經(jīng);;社會(huì)消費(fèi)品零售總額ARIMA模型的建立及預(yù)測(cè)[J];科技廣場(chǎng);2006年10期

10 張華初;林洪;;我國(guó)社會(huì)消費(fèi)品零售額ARIMA預(yù)測(cè)模型[J];統(tǒng)計(jì)研究;2006年07期

相關(guān)博士學(xué)位論文 前1條

1 王娟;基于消費(fèi)者行為的零售業(yè)態(tài)演進(jìn)研究[D];中南大學(xué);2012年

相關(guān)碩士學(xué)位論文 前3條

1 呂曉峰;我國(guó)電子商務(wù)各細(xì)分市場(chǎng)規(guī)模與社會(huì)消費(fèi)品零售總額之間的關(guān)系研究[D];云南財(cái)經(jīng)大學(xué);2015年

2 李勛龍;我國(guó)建筑業(yè)影響因素的實(shí)證分析[D];華中師范大學(xué);2014年

3 萬(wàn)麗娟;基于灰色模型和ARMA模型的我國(guó)社會(huì)消費(fèi)品零售總額的研究[D];中南大學(xué);2009年

，

本文編號(hào)：2091189