- 规划与建设
- 文章编号：1009-6000(2026)03-0073-09
- 中图分类号：TU984 文献标识码：B
- Doi：10.3969/j.issn.1009-6000.2026.03.010
- 项目基金：国家重点研发计划项目课题“城市建成环境‘人—车—物’活动流时空建模分析与刻画”（2024YFB3908603）。
- 作者简介：袁存宇，武汉大学城市设计学院，硕士研究生；
张霞，通信作者，武汉大学城市设计学院，教授。
- 基于自然语言处理的商业建成环境公众情绪影响效应研究——以武汉 9 座商业项目为例
- A Case Study on the Public Sentiment Effects of Commercial Built Environment Based on Natural Language Processing: Nine Commercial Projects in Wuhan
- 袁存宇 张霞
- YUAN Cunyu ZHANG Xia
- 摘要：
商业建成环境的感知与评价对未来商业空间规划、设计和运营具有重要的指导意义。研究以武汉市的 9 座商业项目为例，通过结合自然语言处理技术、RF-SHAP 机器学习解释模型和情感流变分析，对社交媒体文本进行深度语义挖掘。通过自下而上的研究视角，捕捉了公众对商业空间的体验感知，识别出公众感知的周期性变化和影响情感的关键建成环境要素，实现了多时空维度的精细化评价。RF-SHAP 结果揭示了商业环境与公众感知的复杂关系。研究深入分析了建成环境要素如何影响公众情感，其多维度的见解为商业建成环境的定量时空监测和响应公众需求变化提供了新思路。 - 关键词：
社交媒体数据；自然语言处理；机器学习；商业建成环境；公众感知 - Abstract: The perception and evaluation of commercial built environment hold significant implications for future commercial space planning, design, and operation. This study examines nine commercial projects in Wuhan, employing a combination of natural language processing techniques, RF-SHAP machine learning interpretation models, and sentiment flow analysis to conduct in-depth semantic mining of social media texts. Through a bottom-up research perspective, this study captures public experiential perceptions of commercial spaces, identifies cyclical changes in public perception, and recognizes key built environment elements that influence public sentiment, achieving a multi-dimensional spatiotemporal evaluation. The RF-SHAP results reveal the complex relationship between the commercial environment and public perception. This study provides an in-depth analysis of how built environment factors influence public sentiment, offering multidimensional insights that contribute to the quantitative spatiotemporal monitoring of commercial built environments and responding to evolving public needs.
- Key words: social media data; natural language processing; machine learning; commercial built environment; public perception