统计与数据科学学院知识库

  贫困是世界性问题，其动态性、普遍性和易发性等多种特征，导致地区经济发展迟缓、社会发展不稳定，成为全面提高人民生活水平的阻碍。我国作为世界上最大的发展中国家，始终在与消除贫困做斗争。当前，脱贫攻坚已取得重大胜利，但是脱贫摘帽并不代表贫困问题走向终结，刚刚越过贫困标准线的长期贫困人口依然面临返贫的风险，如何让该类人群和地区实现稳定脱贫成为当前帮扶工作的重点。

  基于上述问题，本文以原扶贫开发工作重点县甘肃省秦安县为例，以可持续生计理论为基础，构建相应的指标体系对脱贫农户的脱贫稳定性进行分析。首先构建脱贫稳定性指标体系计算农户的脱贫稳定性指数，运用模糊聚类方法对秦安脱贫农户的脱贫稳定性等级进行科学划分，运用空间自相关对秦安县脱贫稳定性空间上的相关关系进行探究，同时尝试构建返贫预警模型。以期寻找提升秦安县脱贫农户脱贫稳定性的有效途径。

  结果表明，秦安县部分农户依然存在脱贫稳定性不强的问题。通过模糊聚类对脱贫稳定性进行测度分析，发现处于A1（重度不稳定）类型的农户占14.69%，稳定性还有待提高。运用空间自相关研究脱贫稳定性的空间相关关系发现，秦安县脱贫稳定性在空间上具有一定的相关关系，其中王铺镇，魏店镇和安伏镇局部空间自相关呈现低—低相关性，需要重点整治。通过多分类SVM模型构建返贫预警机制，结果表明模型具有良好的效果和一定的可行性。因此，从提升自然资源利用效率，提升人力资本质量两方面，提出提升秦安县脱贫稳定性的对策建议。

  Poverty is a worldwide problem, which is characterized by its dynamic nature, universality and vulnerability, leading to sluggish regional economic development and unstable social development, and becoming an obstacle to comprehensively improving people's living standards. As the largest developing country in the world, China has been struggling with poverty eradication. At present, major successes have been achieved in poverty alleviation, but poverty eradication does not mean the end of poverty. Long-term poor people who have just crossed the poverty line are still facing the risk of returning to poverty. How to stabilize these groups and regions to achieve poverty alleviation has become the focus of current assistance work.

    Based on this, this thesis takes Qin 'an County in Gansu Province as an example. Based on the sustainable livelihood theory, the corresponding index system is constructed to measure and analyze the stability of poverty alleviation of peasant households. Firstly, the index system and index calculation of poverty alleviation stability were constructed. Fuzzy clustering method was used to scientifically analyze the poverty alleviation stability of farmers in Qin 'an County. Spatial autocorrelation was used to explore the spatial correlation of poverty alleviation stability in Qin 'an County. In order to find an effective way to improve the stability of poverty alleviation farmers in Qin 'an County.

    The results show that some farmers in Qin 'an county still have the problem of poor stability in poverty alleviation. Through the measure analysis of the stability of poverty alleviation by fuzzy clustering, it is found that the number of households in A1 (severe instability) type accounts for 14.69%, and the stability needs to be improved. Using spatial autocorrelation to study the spatial correlation of poverty alleviation stability, it is found that the poverty alleviation stability of Qin 'an County presents spatial agglomeration, among which the local spatial autocorrelation of Wangpu town, Weidian Town and Anfu Town shows low-low correlation, which needs to be focused on rectification. The multi-classification SVM model was used to construct the early-warning mechanism for returning to poverty. The results show that the model has good effect and certain feasibility. Therefore, from two aspects of improving the utilization efficiency of natural resources and improving the quality of human capital, countermeasures and suggestions are put forward to improve the stability of poverty alleviation in Qin 'an County.