基于偏序集的农村多维贫困评价研究

Research on the evaluation of rural multidimensional poverty on the perspective of poset

农村多维贫困程度评价直接影响扶贫政策的针对性和有效性,而指标权重设置是制约多维贫困评价准确性和说服力的核心问题。针对现有赋权方法得到的结果存在较大差异的缺陷以及具体准则权重不可知性的属性,提出一种新的基于偏序集的多维贫困评价方法,并利用中国居民营养和健康调查(CHNS)的数据对12个省和直辖市的多维贫困指数进行比较研究。结果发现,由指标重要性差异将12个区域按多维贫困程度划分为五个层次,上海、北京属于多维贫困程度最低层次,河南属于多维贫困程度最高层次,其它区域介于二者之间。基于偏序集的多维贫困评价不仅解决了多维贫困评价指标集结过程中权重难以精确测量的问题,而且指标聚合后的多维贫困排序结果稳健、效果良好。

Evaluation of rural multidimensional poverty directly affects the pertinence and effectiveness of poverty alleviation policy.Weight setting is the core problem that restricts the accuracy and persuasion of multidimensional poverty index.In this paper,a new multidimensional poverty index synthesis method based on partial order set was proposed.Based on the data of China Health and Nutrition Survey(CHNS),a comparative study was conducted on the multidimensional poverty level of twelve provinces and municipalities directly under the central government.The results showed that the twelve regions can be divided into five levels by the significant difference of indicators.Shanghai and Beijing belong to the first level,while Henan belongs to the fifth level,while the other three levels are in between.The multidimensional poverty based on poset can not only solve the problem that the weight of indicators are difficult to accurately measure,but also get an robust and effective result.