【作者】 魏维轩； 王南；

【机构】 上海济致建筑规划设计有限公司极致数据工作室； 同济大学； 南京工业大学建筑学院城乡规划系；

【摘要】 在全域旅游、互联网与人工智能飞速发展的背景下,旅游路径常以旅游目的地城市为主、以核心景区为辅,旅游者的出行常受到网评体验的影响。在各网站海量网评中,采用机器学习方法挖掘数据,构建旅游目的地城市评价与分类规则模型,是以风景园林学科为主的旅游规划与大数据应用之间的交叉探索,也是更科学性的计算机辅助旅游目的地筛选与智慧景区规划的决策参考。主要内容包括:(1)基于空间数据挖掘与机器学习算法理论,定义评分、评价热度、评热比,提出旅游目的地评价决策理论;(2)选择旅游网站城市评价端口,挖掘热评数据,构建评分梯度优化与排序聚类模型、热度聚类评分与排序规则模型及旅游目的地城市分类规则模型,总结目的地城市评热比与城市风景资源分类规则的应用场景;(3)以Python网络数据挖掘脚本于旅游门户网站抓取中国342个城市中共计3118293个旅游目的地评价数据,表明:网络评价前3位的旅游目的地城市分别为那曲(西藏)、博尔塔拉(新疆)、克孜勒苏柯尔克孜(新疆)。网络热度前3位的旅游目的地城市依次为北京、上海、厦门,评热比前3位的旅游目的地城市分别为那曲(西藏)、阿里(西藏)、阿勒泰(新疆)。结合地形的旅游目的地城市分类规则表明:高评价目的地集中于西北、西南、华南、东北北部与长三角地区,地形景区受欢迎度依次为山脉、高原、盆地、平原,游客较人文景点而言更偏爱自然风景旅游地。更多还原

【Abstract】 With the rapid development of holistic tourism,internet and artificial Intelligence,tourist route often takes the tourist destination city as the center,with the core scenic spot as the supplement,the travel of the tourist is often influenced by the internet assessment.In that large-scale web review of each website,using machine learning method to conduct data mining,assessment and grouping model of tourism destination is established,which is a cross-exploration between tourism planning and large data application based on landscape architecture.It is also a scientific computer-aided tourist destination screening and intelligent scenic spot planning decision-making reference.The main contents include:(1) based on the theory of spatial data mining and machine learning algorithm,the definition of assessment value,assessment quantity,assessment value vs.assessment quantity ratio,and the decision theory of tourism destination evaluation are proposed;(2) by selecting the tourism website city evaluation port,mining the thermal assessment data,building graded gradient optimization,sorting clustering,the model of popular clustering grading and collation,the model of urban classification rules of tourist destinations,the application scenarios of the classification rules on urban landscape resources as well as the evaluation ratio of destination cities are summarized;(3) using Python network data Mining script to crawl the 342 cities in China,a total of 3,118,293 tourism destinations evaluation data,the results show that the first three top tourist destinations in the network evaluation are Nakchu Prefecture of Tibet,Bortala Mongol Autonomous Prefecture of Xinjiang,KizilsuKirghiz Autonomous Prefecture of Xinjiang;the network of the top three tourist destinations are Beijing,Shanghai,Xiamen;the top three popular tourism destination cities in the website are respectively Nakchu Prefecture of Tibet,Ngari Prefecture of Tibet,Altay of Xinjiang;the urban classification rules of the tourist destinations combined with topography show that:the high evaluation destinations are concentrated in the northwest,Southwest,south,northeast and Yangtze River delta areas,and the popularity of the terrain scenic area is the mountain,plateau,basin and plain,and tourists prefer the natural scenery tourism to the scenic spots.更多还原