防止返贫动态监测及监测状态变换成因探究——基于多维贝叶斯网络分类器模型的分析  

Dynamic Monitoring to Prevent the Return to Poverty and Exploring the Causes of the Change in Monitoring Status:Analysis Based on the Multi-dimensional Bayesian Network Classifier Model

在线阅读下载全文

作  者:平卫英[1,2] 郭玉帑 黄斐 PING Weiying;GUO Yutang;HUANG Fei(School of Statistics and Data Science,Jiangxi University of Finance and Economics;Applied Statistics Research Center,Jiangxi University of Finance and Economics)

机构地区:[1]江西财经大学统计与数据科学学院 [2]江西财经大学应用统计研究中心

出  处:《中国农村经济》2025年第2期86-109,共24页Chinese Rural Economy

基  金:国家社会科学基金重大项目“后扶贫时代中国城乡相对贫困统计测度与治理机制研究”(编号:20&ZD131);江西省宣传思想文化领域高层次人才专题项目“中国农村低收入群体识别、监测与治理机制研究”(编号:23ZXRC13);江西省研究生创新专项资金项目“防止返贫动态监测及动态转换成因研究”(编号:YC2024-B166)。

摘  要:立足现阶段防止返贫动态监测和帮扶的迫切需要,本文基于江西省2021-2023年的三期追踪调查数据,借助多维贝叶斯网络分类器模型预测农户返贫致贫风险。在此基础上,本文采用标准下行半偏差调整后的概率测量农户多维贫困脆弱性水平,更有效地甄别2020年后返贫治理的监测对象。进一步,本文利用有序Logit模型考察了农户防止返贫监测状态动态变换的风险诱因。研究表明:多维贝叶斯网络分类器模型能够较好地估算农户返贫致贫风险,对农户监测状态的整体预测精度达90%以上。而且,该模型可以根据实时数据动态更新农户的监测状态,为制定帮扶政策提供坚实的决策支持。人力资本和金融资本的减少是农户难以摆脱返贫风险状态的重要因素,而参与社会组织、改善住房条件等能够有效降低返贫致贫风险。此外,本文对防止返贫监测和完善帮扶机制提出了相关建议。Based on the urgent need for dynamic monitoring and assistance to prevent poverty return in China at the current stage,it is of great significance to build an efficient and reliable monitoring and early warning mechanism for the risk of returning to poverty to prevent a large-scale return to poverty and consolidate and expand the achievements of poverty alleviation.Based on the data of the longitudinal survey in Jiangxi Province from 2021 to 2023,this paper uses a multi-dimensional Bayesian network classifier model to predict the risk of rural households returning to poverty.On this basis,this paper measures the multi-dimensional poverty vulnerability level of rural households through the standard downward half-deviation adjusted probability,thus more effectively identifying the monitoring objects of govermance of returning to poverty after 2020.To explore the frequency of monitoring state transformation,the evolution law of instability degree,and influencing factors,the risk triggers of the dynamic transformation of the monitoring state of the farmers to prevent the return of poverty are analyzed using the ordered Logit model.This paper finds that the overall prediction accuracy of the multidimensional Bayesian network classifier model based on the vulnerability theory for the monitoring status of farmers is more than 90%.In addition,the overall multidimensional poverty vulnerability level of rural households is below 10%,covering a wider range of vulnerable groups than those who are actually in distress.Further research shows that the reduction in human capital and financial capital is an important factor for rural households to get rid of the risk of returming to poverty while participating in social organizations and improving housing conditions can effectively reduce the risk of retuming to poverty.Based on the above conclusions,this paper proposes the following suggestions.First,it is necessary to improve the monitoring mechanism for preventing the return to poverty based on muli-source fusion data and big

关 键 词:防止返贫动态监测 多维贝叶斯网络分类器 下行风险 动态变换 

分 类 号:F328[经济管理—产业经济]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象