[1]许芗斌,谢从迈.基于大语言模型的历史名园感知特征研究——以重庆市鹅岭公园为例[J].中国名城,2026,40(4):46-54.[doi:10.19924/j.cnki.1674-4144.2026.004.006]
XU Xiangbin,XIE Congmai.Research on the Perceptions of Historical Gardens Based on Large Language Models: A Case Study of E’ ling Park, Chongqing[J].China Ancient City,2026,40(4):46-54.[doi:10.19924/j.cnki.1674-4144.2026.004.006]
点击复制
基于大语言模型的历史名园感知特征研究——以重庆市鹅岭公园为例()
中国名城[ISSN:1674-4144/CN:32-1793/GO]
- 卷:
-
40
- 期数:
-
2026年4期
- 页码:
-
46-54
- 栏目:
-
- 出版日期:
-
2026-04-05
文章信息/Info
- Title:
-
Research on the Perceptions of Historical Gardens Based on Large Language Models: A Case Study of E’ ling Park, Chongqing
- 文章编号:
-
1674-4144（2026）004-0046-09
- 作者:
-
许芗斌; 谢从迈
-
许芗斌，重庆大学建筑城规学院风景园林系副系主任，副教授，重庆大学山地城镇建设与新技术教育部重点实验室成员，博士；谢从迈（通信作者），重庆大学建筑城规学院硕士研究生，重庆大学山地城镇建设与新技术教育部重点实验室成员。
- Author(s):
-
XU Xiangbin; XIE Congmai
-
-
- 关键词:
-
历史名园; 大语言模型; 公众感知; 生态系统文化服务; 重庆市
- Keywords:
-
historical gardens; large language model; public perception; cultural ecosystem services; Chongqing City
- 分类号:
-
TU985
- DOI:
-
10.19924/j.cnki.1674-4144.2026.004.006
- 文献标志码:
-
A
- 摘要:
-
历史名园作为承载城市记忆、文化象征与审美体验的重要场所，在延续文脉、提升城市品质与满足公众休闲需求方面具有重要意义，但传统感知研究受限于样本规模与语义解析能力，难以系统呈现其真实体验。以重庆鹅岭公园为例，基于大众点评评论文本，以生态系统文化服务为理论视角，构建结合大语言模型的感知解析框架，实现对感知维度、情绪倾向及关键要素的自动识别与量化。结果表明，自然美学、地方认同与休闲活动构成主导感知维度，可达性与配套供给是影响体验的主要短板；心理健康与艺术灵感虽感知频次较低，但满意度显著偏高，具有潜在提升价值。在要素层面，公园的季相、地标意象与观景眺望是核心优势，停车、路线组织与服务管理为负面反馈集中点，在摄影与精神疗愈等方面展现出独特价值。基于上述发现，提出以巩固主导要素、补齐短板要素与激活潜力要素为路径的感知友好型活化更新策略，为历史城市空间的体验优化与精细化治理提供新的方法参考。
- Abstract:
-
Historical gardens, as key sites that embody urban memory, cultural symbolism, and aesthetic experience, play an essential role in sustaining historical continuity, enhancing urban quality, and meeting public leisure needs. However, conventional perception studies are constrained by limited sample sizes and shallow semantic interpretation, making it difficult to systematically capture visitors’ living experiences. Taking E’ ling Park in Chongqing as a case study, this research draws on user-generated reviews from Dianping and adopts the theoretical lens of Cultural Ecosystem Services (CES) to develop a perception analysis framework enhanced by large language models (LLMs). This framework enables the automated identification and quantification of perception dimensions, emotional valence, and key experiential elements. The results indicate that natural aesthetics, place identity, and leisure activities constitute the dominant perception dimensions, whereas accessibility and supporting facilities emerge as major experiential shortcomings. Although psychological well-being and artistic inspiration appear less frequently, they exhibit notably high satisfaction levels, revealing substantial latent value. At the element level, seasonal scenery, landmark imagery, and scenic viewing represent core experiential strengths, while parking, route organization, and service management are the main sources of negative feedback. The historical garden also demonstrates distinctive experiential value in photography and emotional restoration. Building on these findings, the study proposes a perception-oriented activation and renewal strategy centered on reinforcing dominant elements, addressing weak elements, and leveraging high-potential elements. This approach offers methodological insights for enhancing visitor experience and advancing refined governance in historical urban spaces.
更新日期/Last Update:
2026-06-02