摘要
失地农户作为城镇化过程中因征地拆迁而产生的特殊群体,其生计策略选择一直受到学者们的关注。为此,构建失地农户“生计资本-适应能力-制度环境-生计策略选择”的综合分析框架,利用西安市的调查数据,采用网络分析法(ANP),分析生计资本、适应能力、制度环境对失地农户生计策略选择的影响,并构建失地农户生计策略选择模型,在此基础上利用支持向量机(SVM)构建其应用模型。研究结果显示:(1)失地农户可选择的生计策略包括以农业种植活动为主、以个体经营为主、以外出打工为主和以获取工资性收入为主4种类型。(2)具有不同资本禀赋的家庭在生计策略选择方面存在差异,通过ANP和SVM构建的模型可为失地农户的生计策略选择提供指导。
As a special group arising from land acquisition and house demolition in the process of urbanization,farmers who have lost their land have always been concerned about their livelihood strategy choices.To this end,a comprehensive analysis framework of“livelihood capital adaptive capacity-institutional environment livelihood strategy selection”for land-losing farmers is constructed.Using the survey data of Xi’an city,summarize the existing livelihood strategies of land-losing farmers,analyze the impact of livelihood capital,adaptability,and institutional environment on the livelihood strategy choices of farmers facing land-losing impacts through network analysis(ANP),and construct model of land-losing farmers livelihood strategy choice,and finally use the support vector machine(SVM)to build its application model on this basis.The results of the study show:(1)The livelihood strategies of land-losing farmers include mainly agricultural planting activities,self-employed operations,migrant employment and getting wage income.(2)Families with different capital endowments have differences in the choice of livelihood strategies.The model constructed through ANP and SVM can provide guidance for the choice of land-losing farmers’livelihood strategies,and the prediction results of SVM and the degree of fit of the selection plan obtained by ANP reach 90.3%.
作者
周恩毅
聂思言
ZHOU Enyi;NIE Siyan(School of Public Administration,Xi’an University of Architecture&Technology,Xi’an 710055,China)
出处
《西北农林科技大学学报(社会科学版)》
CSSCI
北大核心
2021年第6期126-137,共12页
Journal of Northwest A&F University(Social Science Edition)
基金
教育部人文社会科学研究项目(20XJA630004)
陕西省社会科学基金项目(2019S028)。
关键词
失地农户
生计策略
网络分析
SVM
机器学习
land-losing farmer
livelihood strategy
network analysis
SVM
machine learning