摘要
使用K-means聚类方法对我国农村贫困地区的贫困人口进行聚类,并进一步分析特殊类型贫困地区、集中连片贫困地区的贫困类型结构。结果表明,贫困类型的分布呈现了扶贫对象在区域间分布的不平衡性,各种贫困类型的不同特点和区域分布上的差异从一个视角揭示了收入差距特别是贫困程度差异化的来源。尤其是,少数民族地区的贫困特征和贫困人口比重都要比老区和边境县地区更加突出,而这些地区有着自身独特的特点和性质,尤其需要对少数民族地区贫困背后的形成机制开展更加深入的研究,以便提出针对少数民族地区的因地制宜的扶贫开发措施。连片特困地区的主导贫困类型各不相同,意味着片区扶贫开发需要具有片区针对性的扶贫政策。尽管聚类分析只是一种探索性分析,但是农村贫困人口的聚类仍然为我们定义各种贫困的类型、以及它们在不同区域或特定区域划分之间的内部分布结构提供了非常有价值的信息,并将为进一步的统计推断分析提供基础。
Using the method of K-means clustering, this paper makes the classification poverty population in rural China and thus the analysis of structure of poverty types in areas of special types of poverty and in contiguous poverty areas. The outcomes show that the targeted poor are disproportionately distributed among regions and the features of different types and their regional distribution can be treated as sources of income inequality especially the poverty levels. In particular, poverty characteristics are more notable and the poverty is lager in population in ethnic minority areas than those in old revolutionary base areas and border regions, which implicates that further research is required to explore the hiding mechanism causing poverty in ethnic minority areas so as to put forward poverty alleviation and development measures accommodating to local condition. Also, the leading poverty type is different among contiguous poverty-stricken areas, so that targeted policies are needed. Though clustering is mainly deemed as exploratory analysis, the clustering of rural poverty population still helps to make classifications and definitions of various types of poverty and the internal structure and regional distribution of these poverty types, which can contribute to further statistical inferences and causal analysis.
出处
《中国农业大学学报(社会科学版)》
CSSCI
北大核心
2015年第2期98-109,共12页
Journal of China Agricultural University;Social Sciences
基金
2010年国家社会科学基金重大招标项目"我国特殊类型贫困地区扶贫开发战略研究"(项目号:10zd&025)
"中国人民大学科学研究基金(中央高校基本科研业务费专项资金资助)项目"(项目号:13XNH153)研究成果
国家留学基金委对第一作者在美访问研究的资助(录取文号:留金发[2013]3009)
关键词
农村贫困人口
K-MEANS聚类
特殊类型贫困
连片特困地区
区域分布
Rural poverty population
K-means clustering
Special types of poverty
Contiguous poverty-stricken areas
Geographical distribution