本文選題：偏正態(tài)混合效應(yīng)模型　切入點：碳強度　出處：《杭州電子科技大學(xué)》2017年碩士論文　論文類型：學(xué)位論文

【摘要】：隨著我國經(jīng)濟的高速增長,以及工業(yè)化和城鎮(zhèn)化進程的不斷推進,能源消費總量不斷上升。巨大的能源消耗帶來了二氧化碳的大量排放,環(huán)境問題日益凸顯。而我國已在全球碳排放量排名中位居第一,一方面需要在國際社會承擔(dān)相應(yīng)的碳減排任務(wù),另一方面需要維持經(jīng)濟的快速發(fā)展。因此,提高能源利用效率,降低碳強度,促進我國低碳、可持續(xù)發(fā)展已成為各方學(xué)者研究重點。然而,經(jīng)過顯著性檢驗,我國各地區(qū)碳強度縱向數(shù)據(jù)并不服從正態(tài)分布,若仍以傳統(tǒng)計量經(jīng)濟方法進行建模研究,會使統(tǒng)計分析結(jié)果缺乏穩(wěn)健性。鑒于此,立足于實際碳強度數(shù)據(jù)分布特征,對其構(gòu)建偏正態(tài)分布模型進行統(tǒng)計推斷,并據(jù)此研究碳強度影響因素,為完善我國碳減排政策提供對策與建議顯得尤為重要。首先,本文基于EM算法給出偏正態(tài)混合效應(yīng)模型中未知參數(shù)的極大似然估計。進而,應(yīng)用參數(shù)Bootstrap構(gòu)建偏正態(tài)單項分類模型感興趣參數(shù)的精確檢驗方法。在此基礎(chǔ)上,選取2000至2014年我國各省市自治區(qū)碳強度數(shù)據(jù),驗證其偏正態(tài)分布特征。繼而,構(gòu)建偏正態(tài)混合效應(yīng)模型,研究我國各省市自治區(qū)碳強度變動的主要影響因素。最后,將上述模型參數(shù)估計結(jié)果與正態(tài)混合效應(yīng)模型進行比較,以說明偏正態(tài)混合效應(yīng)模型的優(yōu)良性。研究結(jié)果表明,人均GDP、能耗強度、第二產(chǎn)業(yè)比重、對外貿(mào)易依存度等因素的變化,均會對我國各省份的碳強度產(chǎn)生顯著影響。其中,加大科研投入并促進技術(shù)進步、降低第二產(chǎn)業(yè)比重以及促進對外開放水平等措施,均有利于降低我國各省份碳強度。而通過降低能耗強度來提高能源利用效率,從而降低碳強度是最為直接、有效的方式。同時,我國目前仍處于粗放型的經(jīng)濟發(fā)展中,需加快向集約型的經(jīng)濟發(fā)展方式轉(zhuǎn)變。
[Abstract]:With the rapid economic growth of our country, as well as the continuous progress of industrialization and urbanization, the total amount of energy consumption has been rising. The huge energy consumption has brought a large amount of carbon dioxide emissions. Environmental problems are becoming increasingly prominent. China has already ranked first in the global ranking of carbon emissions. On the one hand, it needs to undertake the corresponding task of reducing carbon emissions in the international community, on the other hand, it needs to maintain rapid economic development. Reducing carbon intensity, promoting low carbon, sustainable development has become the focus of scholars. However, after significant test, the longitudinal data of carbon intensity in different regions of China are not suitable for normal distribution. If the traditional econometric method is still used to model the model, the results of statistical analysis will be lack of robustness. In view of this, based on the distribution characteristics of actual carbon intensity data, the statistical inference is made on the construction of partial normal distribution model. Based on this, it is very important to study the influencing factors of carbon intensity, and to provide countermeasures and suggestions for the improvement of China's carbon emission reduction policy. Firstly, based on EM algorithm, the maximum likelihood estimation of unknown parameters in the model of partial mixing effect is given, and then, the maximum likelihood estimation of the unknown parameters in the model is given based on the EM algorithm. On the basis of the accurate test method of the interested parameters of the biased individual classification model based on the parameter Bootstrap, the carbon intensity data of provinces, cities and autonomous regions of China from 2000 to 2014 are selected to verify the characteristics of partial normal distribution. A partial mixing effect model is constructed to study the main factors affecting carbon intensity change in China's provinces, cities and autonomous regions. Finally, the estimated results of the above model parameters are compared with the normal mixing effect model. The results show that the changes of per capita GDP, energy consumption intensity, secondary industry specific gravity and foreign trade dependence will have a significant impact on the carbon intensity of the provinces in China. Such measures as increasing the investment in scientific research and promoting technological progress, reducing the proportion of secondary industries and promoting the level of opening to the outside world are all conducive to reducing the carbon intensity in various provinces of China. Therefore, reducing carbon intensity is the most direct and effective way. At the same time, our country is still in the extensive economic development, so it is necessary to speed up the transformation to intensive economic development mode.
【學(xué)位授予單位】：杭州電子科技大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：F124;X321

【參考文獻】

相關(guān)期刊論文 前10條

1 Ren Dao YE;Tong Hui WANG;;Inferences in Linear Mixed Models with Skew-normal Random Effects[J];Acta Mathematica Sinica;2015年04期

2 孟生旺;肖展航;;基于偏正態(tài)隨機效應(yīng)模型的信度保費[J];統(tǒng)計研究;2015年01期

3 王明高;;偏正態(tài)分布與偏t正態(tài)分布對保險損失數(shù)據(jù)的擬合分析[J];統(tǒng)計與決策;2014年22期

4 李玲雪;吳劉倉;詹金龍;;缺失偏態(tài)數(shù)據(jù)下聯(lián)合位置與尺度模型的統(tǒng)計推斷[J];統(tǒng)計與信息論壇;2014年03期

5 孫欣;張可蒙;;中國碳排放強度影響因素實證分析[J];統(tǒng)計研究;2014年02期

6 李新運;吳學(xué)錳;馬俏俏;;我國行業(yè)碳排放量測算及影響因素的結(jié)構(gòu)分解分析[J];統(tǒng)計研究;2014年01期

7 吳劉倉;張家茂;邱貽濤;;缺失偏態(tài)數(shù)據(jù)下線性回歸模型的統(tǒng)計推斷[J];統(tǒng)計與信息論壇;2013年09期

8 吳劉倉;馬婷;戴琳;;基于StN分布下聯(lián)合位置與尺度模型的極大似然估計[J];應(yīng)用數(shù)學(xué);2013年03期

9 馬婷;吳劉倉;黃麗;;基于偏正態(tài)分布聯(lián)合位置、尺度與偏度模型的極大似然估計[J];數(shù)理統(tǒng)計與管理;2013年03期

10 徐盈之;徐康寧;胡永舜;;中國制造業(yè)碳排放的驅(qū)動因素及脫鉤效應(yīng)[J];統(tǒng)計研究;2011年07期

相關(guān)博士學(xué)位論文 前1條

1 葉仁道;幾類線性統(tǒng)計模型的估計和檢驗[D];北京工業(yè)大學(xué);2008年

相關(guān)碩士學(xué)位論文 前3條

1 周龍權(quán);基于偏正態(tài)分布的中國能耗強度統(tǒng)計建模研究[D];杭州電子科技大學(xué);2015年

2 齊元方;中國區(qū)域經(jīng)濟增長與碳排放地區(qū)差異斂散性分析[D];湘潭大學(xué);2012年

3 魏雄軍;具有偏正態(tài)隨機擾動項AR模型的統(tǒng)計推斷[D];吉林大學(xué);2009年

，

本文編號：1563125