摘要
社区生活圈规划已成为学术研究与规划实践的热点议题和我国国土空间规划体系创新的重要组成部分。首先,从社区生活圈的概念和内涵界定、空间范围划分方法与设施评价和优化等3个方面对现有规划研究进行综述,并提出社区生活圈规划研究在理论、技术方法和实证方面所面临的挑战。对于居民日常行为的关注是社区生活圈规划的主要特点,因此梳理机器学习方法应用于时空行为研究的前沿与热点主题,包括时空行为预测、时空行为模式挖掘与时空行为和建成环境的非线性关系等。其次,提出机器学习方法在社区生活圈规划研究中的应用框架,以及在社区生活圈物质空间、社区交往生活圈和安全生活圈3个方面的创新议题。最后,以“基于时空行为需求预测的社区生活圈划分方法”及“基于非线性阈值效应的社区生活圈设施配置规划”两个研究案例阐释机器学习方法在社区生活圈物质空间规划研究的应用。
Community life circle planning has become the frontier and hot spot of academic research and planning practice,and it is also an important part of the innovation of China's territory spatial planning system.We first review the existing community life circle planning research from three aspects:the definition and connotation,the spatial scope delineation methods,and the evaluation and optimization of facilities.We conclude the theoretical,methodological and empirical challenges facing by current research.One of the main features of community life circle planning is the focus on residential daily lives.Therefore,we summarize the frontier and hot topics of the application of machine learning methods in the studies of spatiotemporal behavior,including spatiotemporal behavior prediction,space-time behavior patterns mining,and the non-linear relationships of space-time behavior and built environment.Then,we propose the framework of the application of machine learning techniques in community life circle planning research and the innovation topics,including community life circle physical space,community social interaction life circle and community safe life circle.Finally,we take two examples to illustrate the innovative application of machine learning methods in community life circle planning research,which are"community life circle delineation method based on space-time behavior demand prediction"and"community life circle facility planning based on non-linear threshold effect".
作者
张文佳
李春江
罗雪瑶
柴彦威
ZHANG Wenjia;LI Chunjiang;LUO Xueyao;CHAI Yanwei
出处
《上海城市规划》
北大核心
2021年第4期59-65,共7页
Shanghai Urban Planning Review
基金
国家自然科学基金项目“中国城市社区生活圈的新时间地理学研究”(编号42071203)
国家自然科学基金青年项目“机器学习算法辅助下城市居民多尺度移动行为决策过程与空间优化研究”(编号41801158)资助。
关键词
社区生活圈
机器学习
时空行为
决策树
非线性效应
community life circle
machine learning
space-time behavior
decision tree
non-linear effect
