利用ROST-CM6软件的文本挖掘技术，并基于大众点评网、携程旅行网上海珠广场的评价文本数据，分析广州市民对海珠广场的形象感知情况。通过对高频词、语义结构网络和文本情绪进行分析，并结合问卷调查进一步论证，得到海珠广场在市民心中已有交通便捷、文化氛围浓郁2个主要的形象特征；且其获得的大众评价较高，积极情绪评论占比最多。然而，海珠广场在人群活动类型和环境绿化方面仍有优化空间，应构建更多空间形式来丰富活动类型，并通过改善绿化环境打造特色景观。

This study uses the text mining technique of ROST-CM6 software to analyze the image perception of Guangzhou citizens on Haizhu Square based on the evaluation text data of Dianping and Ctrip. By analyzing the high-frequency words, semantic structure network and text emotion, and further demonstrating with questionnaire survey, it is obtained that Haizhu Square has two main image characteristics in the public's mind: convenient transportation and rich cultural atmosphere. And it receives high public ratings, with the highest percentage of positive sentiment comments. However, Haizhu Square needs to be optimized in terms of the types of crowd activities and environmental greening. More spatial forms should be constructed to enrich the activity types and the environment greening should be improved to create characteristic landscapes.

TU986

国家自然科学基金项目（52008114）；教育部人文社会科学研究青年基金项目（20YJCZH186）；广东省哲学社会科学规划2023年度青年项目（GD23YSH04）