Sichuan Province Research On Dynamic Identification And Early Warning Models Of Poverty Return

After 2020,China will have historically eradicated absolute poverty and successfully addressed widespread regional poverty thanks to the execution of precise poverty alleviation programs,and a "new age of poverty alleviation" would have begun.It is clear that China faces more pressures at the historical crossroads of the "two hundred years" aim as the new era of socialism with Chinese features necessitates an effective linkage between the eradication of poverty and rural revitalization.Establishing an effective early warning system for returning to poverty,identifying people who have left poverty at different levels,and offering individualized assistance are particularly crucial in order to avoid the phenomenon of returning to poverty from undermining the results of poverty eradication and to ensure that China’s poverty alleviation and development work is carried out steadily and with high quality.Poor places are prone to fall into the cycle of "helping the poor,getting out of poverty,and returning to poverty" because poverty is by nature a dynamic reality and the process of returning to it is contingent and repeating.Households that have escaped poverty and non-poor households are very likely to return to poverty due to the detrimental effects of natural conditions,resource endowments,and endogenous dynamics,and poverty alleviation projects targeting through regional objectives are unable to completely cover the issue.The majority of current early warning models for the relapse into poverty are based on logistic regression analysis,and this analysis is predicated on premises that are likely to bias the early warning of the relapse into poverty due to incomplete coverage and object exposure,leading to poor identification accuracy.This leads this paper to further investigate on the basis of prior research,choose a more accurate random forest model to predict households returning to poverty,expand the current poverty return indicator system in light of the actual situation in Sichuan Province,establish a more streamlined and efficient early warning model for identifying and warning households returning to poverty,and propose targeted and effective policy recommendations through clust.82,000 of the more than 6 million people who have been lifted out of poverty in Sichuan Province are at risk of going back into it,and there are also a lot more marginal households that are vulnerable to poverty.As a result,the province is in an unstable condition of poverty eradication.This paper chooses Sichuan Province as the research object,combines theories related to dynamic poverty,multidimensional poverty,and early warning of return to poverty,conducts a thorough and methodical analysis of the issue of return to poverty in Sichuan,and explores the construction of an early warning model for the identification of return to poverty.The first step was to build a system of early warning indicators for the detection of the return of poverty in Sichuan Province.This system included sixteen evaluation indicators,five dimensions of economic,medical and health,education,assets,and living standards,and weights were established using the entropy weighting method in order to choose the early warning indicators with the highest information entropy.The optimal early warning indicator system for poverty return could achieve a prediction accuracy of90.82% through the use of the CFPS data,and a random forest model was then used to screen the early warning indicators for poverty return identification.Through this identification system,households at risk of returning to poverty could be accurately warned.The main potential causes of poverty return and the structure of the main types of households returning to poverty in Sichuan Province are finally further analysed by clustering the random forest predicted household data through K-means clustering analysis,providing a basis for further statistical inference and analysis of poverty alleviation work by classifying various types of poverty return for early warning,in order to establish a clear scientific early warning model.The study’s findings lead to the following interpretations:(1)The early warning function of the mechanism can be maximized by scientific design of early warning indicators for poverty return detection.A multidimensional quantitative examination of environmental,economic,health,and educational dimensions can offer a more thorough and in-depth knowledge of the issue of poverty return than singledimensional evaluation methods.(2)The main element influencing a return to poverty is education.An early warning evaluation model for the recurrence of poverty was built using the random forest approach and K-means cluster analysis,and it was discovered that Sichuan Province had the highest risk due to schooling.(3)Innovative sustainable approaches for managing poverty are urgently needed.The incidence of returning to poverty is primarily concentrated in the categories of years of education,medical spending,Engel’s coefficient,and schooling rate of school-age children,which also shows that Sichuan Province’s household development is generally low in these areas.(4)Policy recommendations for combining poverty eradication achievements and building a reliable,scientific monitoring system for the reemergence of poverty First and foremost,it is important to continuously enhance the early warning system for preventing poverty return.Next,it is important to employ big data analysis techniques like random forest and K-means clustering to increase the accuracy of assistance measures.