基于因子分析的全国各省市综合实力的评价
发布时间：2018-09-03 06:25
【摘要】：当今社会,全国各省市的经济均在快速发展,对各省市的综合实力的研究成为重要课题。若能找到一种评估各省市综合实力的有效方法,在排名过程中确定影响排名的主要因素,针对该因素进行大力建设,将大大提升全国的综合实力。本文主要工作如下：1.主要介绍了聚类分析的判别统计量,以及系统聚类法和动态聚类法。系统聚类法具有根据谱系图确定分类数目及成员的优点,故本文使用系统聚类法。为避免变量的量纲影响,本文采用加权的欧式距离。2.详细介绍了因子分析中的因子模型并求解参数,为了更好的解释公共因子和评价省市综合实力,进行了方差的旋转和因子得分的求解。3.利用国家统计网收集并整理的2012年度全国31个省市的八个经济指标进行了实验。首先,应用聚类分析方法将31个省市分成四类,目的是与最终排名作对比。然后,通过计算KMO和Bartlett球形检验值验证数据适合应用因子分析方法处理；最后,应用因子分析方法提取公共因子并计算公共因子的得分,利用该得分评价各省市的综合实力。实验结果证明,位列前三的分别是广东省、江苏省和山东省,与聚类的分类结果大体一致,影响排名的主要因素是金融业增加值、第三产业增加值和住宿餐饮业增加值,与实际情况相符合。
[Abstract]:Nowadays, the economy of all provinces and cities in China is developing rapidly, so the research on the comprehensive strength of each province and city becomes an important subject. If we can find an effective way to evaluate the comprehensive strength of each province and city, determine the main factors that affect the ranking in the process of ranking, and make great efforts to construct this factor, we will greatly enhance the comprehensive strength of the whole country. The main work of this paper is as follows: 1. This paper mainly introduces the discriminant statistics of cluster analysis, as well as the systematic clustering method and dynamic clustering method. The systematic clustering method has the advantage of determining the number and members of the taxonomy according to the pedigree diagram, so the systematic clustering method is used in this paper. In order to avoid the dimensionality influence of variables, the weighted Euclidean distance. 2. The factor model in factor analysis is introduced in detail and the parameters are solved. In order to better explain common factors and evaluate the comprehensive strength of provinces and cities, the rotation of variance and the calculation of factor score are carried out. Eight economic indicators collected and collated by the National Statistical Network in 2012 from 31 provinces and cities in China were experimented with. First, 31 provinces and cities are divided into four categories by cluster analysis method, and the purpose is to compare with the final ranking. Then, the KMO and Bartlett spherical test data are calculated to verify that the data is suitable to be processed by factor analysis. Finally, the factor analysis method is used to extract common factors and calculate the scores of common factors, which is used to evaluate the comprehensive strength of each province and city. The experimental results show that Guangdong Province, Jiangsu Province and Shandong Province are the top three, and the classification results are consistent with the clustering results. The main factors affecting the ranking are the added value of the financial industry, the added value of the tertiary industry and the added value of the hotel and catering industry. In line with the actual situation.
【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F127;F224
[Abstract]:Nowadays, the economy of all provinces and cities in China is developing rapidly, so the research on the comprehensive strength of each province and city becomes an important subject. If we can find an effective way to evaluate the comprehensive strength of each province and city, determine the main factors that affect the ranking in the process of ranking, and make great efforts to construct this factor, we will greatly enhance the comprehensive strength of the whole country. The main work of this paper is as follows: 1. This paper mainly introduces the discriminant statistics of cluster analysis, as well as the systematic clustering method and dynamic clustering method. The systematic clustering method has the advantage of determining the number and members of the taxonomy according to the pedigree diagram, so the systematic clustering method is used in this paper. In order to avoid the dimensionality influence of variables, the weighted Euclidean distance. 2. The factor model in factor analysis is introduced in detail and the parameters are solved. In order to better explain common factors and evaluate the comprehensive strength of provinces and cities, the rotation of variance and the calculation of factor score are carried out. Eight economic indicators collected and collated by the National Statistical Network in 2012 from 31 provinces and cities in China were experimented with. First, 31 provinces and cities are divided into four categories by cluster analysis method, and the purpose is to compare with the final ranking. Then, the KMO and Bartlett spherical test data are calculated to verify that the data is suitable to be processed by factor analysis. Finally, the factor analysis method is used to extract common factors and calculate the scores of common factors, which is used to evaluate the comprehensive strength of each province and city. The experimental results show that Guangdong Province, Jiangsu Province and Shandong Province are the top three, and the classification results are consistent with the clustering results. The main factors affecting the ranking are the added value of the financial industry, the added value of the tertiary industry and the added value of the hotel and catering industry. In line with the actual situation.
【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F127;F224
【相似文献】
相关期刊论文 前10条
1 王敏;;基于因子分析的济南市软件产业竞争力评价[J];山东行政学院.山东省经济管理干部学院学报;2010年06期
2 蔺全录;孟毅;;基于因子分析法的中小化工企业核心竞争力评价研究[J];企业导报;2013年22期
3 翟航;周俊;李世梅;张丹;辛欣;;因子分析法在污染源解析方面的应用[J];中国环境管理;2014年01期
4 尹子民,罗丽兮;因子分析在企业增长方式评价中的应用[J];数理统计与管理;2000年04期
5 高素英,金浩;因子分析在市场问卷调查分析中的应用[J];河北工业大学学报;2003年05期
6 李升学,康彦彦;我国各地区企业融资差异的多变量因子分析[J];山东农业大学学报(社会科学版);2003年03期
7 杨淼P，
本文编号：2219170
本文链接：https://www.wllwen.com/jingjilunwen/jingjiguanlilunwen/2219170.html
