谁更愿意争当贫困户?——基于机器学习模型的预测与思考

Who Wants to Be Identified as Poor Household?--Based on Machine Learning Model

本文在西部G省大规模实证调研的基础上,通过机器学习模型,稳健预测与检验了谁更愿意争当贫困户。研究发现,思维因素是影响农户是否争当贫困户的关键,文化因素中的村规民约对争当贫困户有重要的影响,心理因素中的脱贫意愿和脱贫信心与争当贫困户有莫大的关系。具体来说,思维模式有偏差、所在村庄没有村规民约、不愿意且没有信心脱贫的人,更愿意争当贫困户。由于争当贫困户的行为,与精准扶贫政策“瞄准偏差”有关,因此要从源头上根治这一现象,一方面应考虑政策制定的政治因素、技术手段和文化现象,另一方面需要结合中国的国情和文化,宏观上优化顶层设计,中观上加强技术监管,微观上强化农户教育。

On the basis of large-scale empirical research in G Province in western China,this paper has applied machine learning model to predict who is more willing to be identified as poor household.The result shows that mentality,the key factor in deciding whether farmers strive to be identified as poor households or not,is followed by the village rules and regulations in cultural factor and the willingness and confidence of poverty alleviation in psychological factor.To be specific,people who have a biased mode of thinking,live in a village without village rules and regulations,and are unwilling and unconfident to get rid of poverty,are more willing to be identified as poor households.Since striving to be identified as poor households is related to the“target deviation”of targeted poverty alleviation policy,the root cause of this phenomenon should be eradicated.On the one hand,the political factors,technical means and cultural phenomena of policy making should be considered;on the other hand,local solutions should be proposed based on China’s national conditions and culture.