摘要
公民共同参与城市治理对于城市的发展具有重要的作用,准确地识别公众参与讨论的主题内容是城市治理中亟待解决的问题。目前,对于公民参与的研究多集中在对平台的评价,对发布的文本内容分析较少。深度学习作为一种以数据为导向的新技术,在自然语言处理相关领域都得到了广泛应用且效果良好。基于政务网站论坛的文本数据,采用大数据和神经网络的方法,设计苏州市公民参与的定量化模型,研究公民在政府网站上的行为,并在不同的时空尺度上研究公民参与度的时空格局演化规律,结合苏州出台的相应政策数据验证方法可靠性。有助于认识一个城市的基层治理能力,评价一个城市可持续性。
Citizen participation in governance plays an important role in urban development.It is,therefore of great interest to accurately identify the topics of resident participation in urban governance.Most of the current research on citizen participation focused on the evaluation of the platform,while the analysis of the published textual content has been ignored if not all.Deep learning has been widely used in natural language processing and shows great potentials in understanding patterns or regularities behind data.In this article,we collect the text data from the government website of Suzhou city,and develop a quantitative model of citizen participation based on deep learning methodology.The model was evaluated by relevant policies of Suzhou city.We also explore the spatio-temporal evolution of citizen participation at multiscale.The method proposed in this article helps evaluate the ability of government management and the sustainability of the city.
作者
沈安南
王亮绪
高峻
SHEN Annan;WANG Liangxu;GAO Jun(School of Environmental and Geographical Sciences,Shanghai Normal University,Shanghai 200234,China;Institute of Urban Study,Shanghai Normal University,Shanghai 200234,China)
出处
《地理信息世界》
2020年第2期44-48,共5页
Geomatics World
基金
国家自然科学基金项目(41730642)资助。
关键词
城市治理
神经网络
地理命名实体识别
时空格局
urban governance
neural network
geographical named entity recognition
spatiotemporal pattern
