【摘要】 通过对山东省某县38个村458户农村居民的走访调查,结果表明农村低保制度虽然一定程度上解决了农村贫困居民经济困难,但扶贫的针对性不足,对特困居民的身份识别不准确,"关系保"、"人情保"等仍然存在。针对这一现状,立足于精准扶贫的理念,在梳理农村低保居民经济特点的基础上,构建了农村低保户身份识别指标体系,并利用主成分分析法对指标体系进行了约简,进而建立了低保户判别的人工神经网络分类模型,同时考虑到低保政策的"兜底"特点,在模型的训练环节上进行了"非对称"处理,以降低将贫困居民判定为非低保户的错误率。实证结果表明,该模型的判别精度和判别效率显著优于标准人工神经网络、多元线性回归等方法,在低保户居民身份识别工作中具有较高的推广应用价值。更多还原

【Abstract】 Through the survey of 458 rural households in a county of Shandong Province,the results show that there are also some shortcomings and problems,especially subsistence object identity is not accurate in the rural minimum subsistence security system. Aiming at the inaccurate identity identification,this paper constructs the rural subsistence object identification index system based on targeted poverty alleviation,and establishes the classified model of rural subsistence based on the Artificial Neural Network( ANN). Empirical results show that the model identification precision and discriminant efficiency is significantly better than that of normal ANN and MRA,which has higher application value in low- income residents’ identity discrimination.更多还原