基于领域词典的留园构成要素情感分析  

Sentiment Clustering of Landscape Constituents of Liuyuan Garden Based on Domain Dictionary

作　　者：刘文龙 黄维 LIU Wen-long;HUANG Wei(Shenzhen International Graduate School,Tsinghua University,Shenzhen 518000,China)

机构地区：[1]清华大学深圳国际研究生院,深圳518000

出　　处：《科学技术与工程》2021年第8期3174-3179,共6页Science Technology and Engineering

摘　　要：在对旅游景点的评论挖掘中常以多景点横向对比为研究切入点,为景点间的横向比较及游人选择景点服务,而较少针对单一景点深入分析,为景点单要素精准提升服务。以留园为例,按照构成元素构建聚类,并基于领域词典进行整体与分要素聚类的情感分析。结果表明,留园中“山石”要素相关的正面情感占比66%,低于分要素平均正面情感78.3%。可见基于园林构成要素聚类分析可帮助精准提取互联网评论情感分析。研究成果对园林等旅游景点管理方优化、品牌形象提升提供了一种易于操作的、更精准的理论与方法。Liuyuan Garden is a well-known ancient Chinese architecture built in 1593 in Ming Dynasty and overhauled in 1876 in Qing Dynasty.As a large-scale classical private garden in China,it features various landscape constituents of halls,corridors,walls,doors,rockeries,pools,flowers,and trees,forming dozens of garden pieces of different sizes.The garden has been listed as a cultural relic of national importance since 1961.In the review mining of tourist attractions,horizontal comparison among multiple scenic spots was often conducted as the research start point in order to provide services for tourists to choose attractions.However,there is a lack of in-depth analysis of a single attraction to provide accurate services.Taking Liuyuan Garden as an example,clustering was constructed according to the constituents,and the sentiment analysis of the whole and sub-element clustering was carried out based on the domain dictionary.Results show that 66% of the positive emotions are related to the“rockeries”elements,which is below 78.3% of the average positive sentiments.It can be seen that cluster analysis based on landscape constituents can help to accurately extract the sentiment analysis of the reviews via the Internet.The results provide a theory basis and a simple and accurate method to optimize the management and promote the brand image of landscape and other tourist attractions.

关 键 词：情感分析 留园 要素聚类 领域词典 

分 类 号：TP391.3[自动化与计算机技术—计算机应用技术]