北京市旅游人群行为情感分析调研报告

发布时间：2018-01-31 06:43

本文关键词： 北京市旅游人群行为特征情感算法　出处：《首都经济贸易大学》2017年硕士论文　论文类型：学位论文

【摘要】：网络旅游平台及大数据技术迅猛发展,旅游平台需借助网络爬虫以及文本处理等技术获知如何完善平台。本报告研究目标就是运用大数据文本分析技术充分挖掘来北京市旅游的人群行为特征如出行时间、客源地结构、出行方式及结伴方式等,同时获取北京市最具热度的前100个景点所对应的文本评论数据,组成具有321550条旅游情感评价的语料库,研究不同行为的旅游人群情感分值差异,从而为完善平台建设提供建议。本报告主要运用以下四种研究方法:第一,文本聚类。将游客评论进行分词并转换成向量,运用K-Means进行聚类,得出旅游人群评价维度。第二,情感分析法。建立情感词典运用情感分析算法来进行文本维度属性情感词的匹配,实现情感分值的量化处理。第三,内容分析法。通过对语料库中景点评论文本进行词频分析,提取与游客情感相关的高频词,细化旅游人群情感评价。第四,对比分析法。对比不同行为的旅游人群情感分值以及评价词语差异。对旅游人群数据进行综合研究得出主要结论:第一,北京对距离近的地区产生更大游玩吸引。第二,出行时间对于公园乐园、古迹遗址以及自然景观类景点的评分影响比较大,情感分值在一年中呈现两端月份低,中间月份高的现象。第三,结伴方式有不同。单独出游、家庭出游、朋友出游、情侣出游以及商务旅行对于景点类型的喜好以及评分有差异。第四,出行方式显个性。选择跟团游、自由行及自驾游人群画像有差异,产品喜好与情感分值评价不同。网络旅游平台可通过以下来进行优化:第一,构建多元评价维度;第二,精确定位推荐旅游产品时间;第三,细化营销推荐人群;第四,优化产品特色服务;第五,加大基础设施投入。
[Abstract]:Internet tourism platform and big data technology are developing rapidly. Tourism platform needs to know how to improve the platform by means of web crawler and text processing technology. The goal of this report is to fully excavate the behavior characteristics of Beijing tourism population by using big data text analysis technology, such as travel time. Between. At the same time, we obtain the text review data of the top 100 scenic spots with the most heat in Beijing, and form a corpus of 321550 tourism emotion evaluation. The study of different behavior of tourism groups emotional score differences, thus providing suggestions for improving the platform construction. This report mainly uses the following four research methods: first. Text clustering. The tourists' comments are partitioned and converted into vectors, and K-Means are used to cluster to get the tourist crowd evaluation dimension. Second. Affective analysis. The establishment of emotion dictionary using emotional analysis algorithm to match the text dimension attributes emotional words, to achieve the quantification of emotional score processing. Third. Content analysis. Through the word frequency analysis of the comment text of scenic spots in the corpus, extract the high-frequency words related to the tourists' emotion, refine the emotional evaluation of the tourist crowd. 4th. Contrastive analysis. Compare the different behavior of tourism groups emotional scores and evaluation of the differences in words. The comprehensive study of the tourist population data draw the main conclusions: first. Beijing has a greater attraction to nearby areas. Second, travel time has a greater impact on park parks, historic sites and natural landscape sites, emotional scores in the year at both ends of the month low. The phenomenon of high in the middle month. Third, there are different ways of getting together. There are differences in the preference and score of individual travel, family trip, friend trip, couple trip and business travel for the type of scenic spot. 4th. Travel style shows personality. Choose with group tour, free travel and self-driving tour crowd portrait differences, product preferences and emotional evaluation is different. Internet tourism platform can be optimized through the following: first. Constructing multiple evaluation dimension; Second, the time of recommending tourism products; Third, refine the marketing recommendation crowd; 4th, optimize product characteristic service; 5th, increase infrastructure investment.
【学位授予单位】：首都经济贸易大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F592.7

【参考文献】