Abstract:The cultural ecosystem services of scenic areas play a crucial role in promoting ecological conservation and enhancing human well-being.The research takes Xishan Scenic Area, Grand View Park, Cuihu Park and Jindian Scenic Area in Kunming as the research objects, and uses social media comment text data and natural language processing techniques to analyze visitor perceptions. A CES perception framework was developed, incorporating social network analysis, Baidu AI sentiment analysis, and importance-performance analysis to explore how visitors perceive CES in these scenic areas. The results indicate that the CES of scenic areas can be classified into six types. Among these, tourists are more likely to perceive the 'historical culture' and 'infrastructure', which are rated with high importance and performance. Conversely, 'social interaction' is perceived less frequently but is considered highly important, with lower performance. Additionally, 'leisure and entertainment' receive lower ratings in terms of perception, importance, and satisfaction. Cuihu Park and Daguan Park exhibit high similarity in CES, while Xishan Scenic Area and Jindian Scenic Area each have their own distinct characteristics. Based on the research results, the CES of the four scenic spots should be improved from four aspects: strengthening historical and cultural heritage and brand building, promoting the digital construction of smart scenic spots, optimizing aesthetic and natural experience from a multi-sensory perspective, and taking into account the carrying capacity to improve the quality of leisure and entertainment.
