基于安徽省大别山区682户农村家庭调查数据，首先采用马田系统从收入、教育、健康、生活条件、资产状况、住房等六个维度精选出导致农村家庭多维贫困的主要因素，然后利用解释结构模型和复杂网络节点重要性分析方法对致贫因素进行分层并测度每一因素对其他因素的影响力，从而识别出导致贫困的深层因素。结果表明：农村家庭贫困多维特征明显，收入贫困已不是主要致贫因素，基础设施建设的不完善是致使农户多维贫困产生的主因，筛选后的因素可以有效区分贫困户与非贫困户；医疗服务可及性、饮用水、做饭燃料和通电情况是农村家庭致贫的深层因素；饮用水、做饭燃料、卫生设施和健康状况对其他致贫因素的影响力较大。

Based on the survey data of 682 rural households in Dabie Mountain Area, Anhui province, the main factors leading to multidimensional poverty in rural households were firstly selected from six dimensions including income, education, health, living conditions, asset status and housing through Mahalanobis-Taguchi System. Then, using the interpretative structure model and the importance analysis method of complex network nodes, the poverty factors were stratified and the influence of each factor on other factors was measured, so as to identify the deep factors leading to poverty. The results show that the multi-dimensional characteristics of rural household poverty are obvious, income poverty is not the main cause of poverty, the imperfect infrastructure construction is the main cause of multi-dimensional poverty, the factors after screening can effectively distinguish poor households from non poor households; medical service accessibility, drinking water, cooking fuel and power supply are the deep-seated factors of rural household poverty; drinking water, cooking fuel, sanitation facilities and health status have great influence on other poverty factors.

F323.8