【目的】通过对城市蓝绿开放空间中使用者使用需求和活动方式的分析,识别城市蓝绿开放空间服务于人的社会介质属性,科学构建评价体系,为丰富“人-蓝绿开放空间-社会”关系的研究及相应管理规划提供参考。【方法】基于文献分析、网络数据抓取及实地调查,从蓝绿开放空间作为城市社会介质的视角出发,构建城市蓝绿开放空间社会介质属性评价模型,并以南京市的3个城市公园为例,运用BP神经网络进行模型验证及指标权重定量分析。【结果】基于样本构建的评价模型包含4项一级指标、12项二级指标,其中一级指标包括空间本底、连接类型、参与方式和空间意义,空间本底(权重0.440)是评价社会介质水平的关键要素;二级指标中影响较大的是自然支持水平(权重0.157),其次是基础设施完善度(权重0.127)、微气候舒适度(权重0.125)和地方认同(权重0.121),其余参数的影响相对较小。分析认为,空间本底质量是吸引使用者进行社会活动的基础,“人-蓝绿开放空间-社会”的亲近关系有助于提升空间的社会介质水平。【结论】采用BP神经网络通过对调查数据的学习,能够实现对城市蓝绿开放空间社会介质属性的评价。将社会学与风景园林领域理论与实践相耦合,从人与空间关系的角度出发,可为城市蓝绿开放空间社会效益的识别与评价提供新的研究思路和方法。

【Objective】Through the analysis of the needs and activity patterns of urban blue-green open space (BGOS) users, the social media attributes of urban BGOS can best serve people were identified. The aim was to construct a scientifically based evaluation system that enhanced the “People-BGOS-Society” relationship and provide a reference for corresponding management planning.【Method】Based on literature research, field surveys, and network data collection, from the perspective of BGOS as the urban social medium, an evaluation model of the social media attributes of urban BGOS was constructed based on four dimensions: space background, connection type, participation mode, and spatial significance. Taking three urban parks in Nanjing as an examples, a back propagation (BP) neural network model was used for model verification and a weighted quantitative analysis.【Result】By studying the survey data, the BP neural network evaluated the social media attributes of urban BGOS. The evaluation model included four first-level indicators and 12 second-level indicators. The first-level indicators included spatial background, connection type, participation mode, and spatial significance, among which spatial background (weight 0.440) was the key factor for evaluating the social media attributes. Among the secondary indicators, the natural support level (weight 0.157) had the greatest impact, followed by infrastructure perfection (weight 0.127), microclimate comfort (weight 0.125), and local identity (weight 0.121), while the remaining parameters showed relatively little impact. The quality of the spatial background was the basis for attracting users to carry out social activities. The close “People-BGOS-Society” relationship helped to improve the spatial level of social media attributes.【Conclusion】This study combined sociology with the field of landscape architecture. From the perspective of the relationship between people and space, the application of a BP neural network model was a novel research concept and methods were developed for the identification and evaluation of the social benefits of BGOS.