基于机器学习的国家公园游客景观偏好研究

Study on Tourists’Landscape Preferences of National Park Based on Machine Learning

为精准探究国家公园游客群体的景观偏好,从而为国家公园相关工作提供有效策略,文章以云南普达措国家公园为研究案例,在对现有文献进行综述后,确定采用社交媒体数据与机器学习联用评估方法,对该公园的游客景观偏好展开研究,以补充以往研究的不足之处。在具体研究过程中,首先以Python 3.11为工具,针对多个网络平台进行普达措国家公园相关信息收集与整理,并应用机器学习方法中的级联多模态分解双线性池化技术,对整理后的数据信息进行深度处理,以提取图像文件中的具体信息要素。而后,针对数据处理结果,应用标签检测方法和多重性质要素检测方法,对各种要素的分布情况进行分析,以此对不同网络平台的游客群体偏好进行初步探究。探究结果显示,在本次分析的微博、抖音和马蜂窝App三个平台中,不同平台用户对普达措国家公园的景观要素存在着不同的偏好,且各具特点。最后,根据总结得到的分析结果,从提升自然环境要素对游客吸引力的数个角度入手,为该公园管理单位的后续工作提供相关建议。

In order to accurately explore the landscape preferences of national park visitors and to provide effective strategies for the related work of national park,the paper takes Yunnan Pudacuo National Park as a research case.After reviewing the existing literature,it determines to use social media data and machine learning to assess the landscape preferences of visitors to the park,in order to supplement the shortcomings of the previous research.In the specific research progress,Python 3.11 is used as a tool to collect and organize the related information of the park from various online platforms.The cascaded multimodal factorized bilinear pooling technique in machine learning is employed to meticulously process the collected data and to extract the specific elements from image files.Subsequently,in terms of the data process results,label detection and multiple feature detection methods are utilized to analyze the distribution of these elements on different online platforms,thus initiating a preliminary exploration on tourist preferences.The research findings reveal that users on Weibo,Tiktok,and Mafengwo App exhibit distinct preferences for the landscape elements of Pudacuo National Park,each displaying unique characteristics.Finally,based on the summarized analysis results,this paper provides relevant suggestions for the subsequent work of the park management unit from several perspectives to improve the attraction of natural environmental elements to the tourists.