Research On The Evaluation And Prediction Of Risk Of Returning To Poverty Based On Machine Learning

China is presently going through a crucial transitional stage as it moves from rural revitalization to poverty alleviation.During this critical period,it is essential to maintain the continuity and predictability of reducing poverty strategies in order to secure the benefits of poverty alleviation.This will help our rural revitalization plan progress steadily and avert any potential slide into poverty.As a result,in the period following poverty alleviation,it is crucial to shift our focus from poverty alleviation to poverty prevention and transform our governance mechanism from being centered on poverty alleviation to focusing on poverty prevention.This necessary step will enable us to consolidate poverty alleviation gains and implement our rural revitalization strategy effectively.To achieve this,it is essential to identify the risk of poverty-stricken households returning to poverty and take preventive measures in advance.This paper investigates the risk of poverty-stricken households relapsing into poverty and provides theoretical support to the government departments for implementing early assistance.Firstly,this paper conducts a comprehensive review of the literature on the risk of relapse into poverty,drawing from relevant policy documents in China,and constructs a scientific and equitable evaluation index system for impoverished households’ risk of returning to poverty from the perspectives of survival ability,development ability,and emergency ability.Secondly,the comprehensive weights are obtained by combining them with the Gini criterion.Thirdly,an evaluation model for the risk of relapse into poverty is constructed using an improved version of the VIKOR method to effectively evaluate the risk of returning to poverty of impoverished households and determine the level of risk.The Qlearning algorithm is introduced to adaptively select and update the strategy to improve the performance of the MPA algorithm,and the enhanced marine predator algorithm is used to optimize the parameters of the model of the Light GBM algorithm in order to build a forecasting models for the risk of going back to poverty using the enhanced Light GBM algorithm.The truth of the evaluation is finally confirmed by contrasting it to different evaluation techniques,and the predictive model is used to predict the future development trend of the risk of returning to poverty of impoverished households,demonstrating its higher prediction accuracy compared with other prediction models.This study provides a theoretical foundation and practical guidance for government departments to implement early assistance to prevent poverty relapse.In summary,this article has analyzed relevant basic theories,constructed evaluation and prediction models,and achieved accurate prediction of the poverty return risk of poverty-stricken households.Through empirical analysis,the poverty return risk of some poverty-stricken households in Zhouzhi County was scientifically evaluated and predicted,comprehensively understanding the changes in poverty return risk of each poverty-stricken household.This provides a theoretical basis for relevant poverty alleviation departments in Zhouzhi County to provide targeted assistance in advance.