【摘要】 从提升人类福祉的角度构建区域社会关系评价指标体系，利用自组织映射神经网络（self-organizing feature map,SOFM）聚类和随机森林（random forest,RF）分类方法计算评价指标的权重.在此基础上，根据2014-2018年的统计数据，采用多目标优化分析（multi objective optimization on the basis of ratio analysis,MOORA）计算中国31个省份的区域社会关系指数（不含港澳台地区），对各省份的区域社会关系发展状况进行了评价.结果显示：1）中国区域社会关系发展呈现波动中上升的趋势，区域社会关系指数由2014年的0.176上升至2018年的0.187;2）区域社会关系指数由高到低依次是东部地区、西部地区、中部地区；3）区域社会关系指数较高的地区，既有较发达地区（如北京、上海、浙江），又有经济相对落后地区（如青海、西藏、宁夏）.本文为区域社会关系的评价提供新的思路和具体可行的计算方法，为改善区域个体社会关系、提升区域人类福祉提供科学依据和决策支持.更多还原

【Abstract】 A regional social relationship-index system was constructed in this work with Self-organizing Feature Map(SOFM) cluster analysis and Random Forest(RF) classification analysis. Statistical data from 2014-2018 was subject to Multi Objective Optimization with Ratio Analysis(MOORA) method, to calculate regional social relationship-index of 31 provinces(municipalities and autonomous regions) in China. Development of regional social relationships in China showed an upward trend in fluctuations,regional social relationship-index increased from 0.176in 2014 to 0.187 in 2018. Ranking of regional social relationship-index from high to low was: the eastern, the western,and the central regions. Regions with high regional social relationship index included relatively developed regions(such as Beijing,Shanghai,Zhejiang) and relatively backward regions(such as Qinghai,Tibet,Ningxia).These findings will help evaluation of regional social relationships,and facilitate improvements in regional individual social relationships and human well-being.更多还原