基于文本大数据的游客旅游意象感知挖掘研究——以重庆市为例

Tourism Image Mining Based on User Generated Big Data:Taking Chongqing as An Example

利用旅游者对重庆市景点的旅游评价大数据,通过基于Word2vec的词向量模型进行语义挖掘,识别城市旅游意象感知要素类型,结合游客情绪表达分析旅游认知意象和情感意象的对应关系,以及意象感知的空间分异。研究发现重庆市的旅游意象分为空间感知、旅游核心吸引物及旅游支持体系这三个层面以及具体十类意象,游客感知以正面积极为主,但也对交通及旅游服务等因素给予较多负面评价。研究表明在利用Word2vec工具处理网络文本基础上对关键词进行聚类,可有效提取意象感知要素,便于分析城市旅游意象结构体系和感知情况,为利用社交媒体大数据实现城市旅游意象感知、指导城市旅游建设提供新的思路。

Big data from social media plays an important role in sensing and evaluating urban tourism image and enhancing urban tourism competitiveness,user generated contents in social network is a valuable source to extract perceived tourism factors and assess the quality of urban tourism construction.With the increase of data volume,traditional content analysis based on word frequency cannot fully exploit the information provided by these data.Natural language processing based on word vector model was introduced into the study of urban tourism image perception.On the basis of data collecting and preprocessing,word vector model is constructed using the whole corpus,and the emotion of every comment are analyzed.Hot tourism spots are identified according to the location of attractions and the number of comments.Keywords extracted by TF-IDF method are then clustered according to the cosine distance of their word vector.Based on cluster results and tourists'emotion,the components of Chongqing's tourism image and elements that caused significant negative emotions were identified.The study finds that Chongqing's tourism image is divided into three aspects:spatial perception,core attractions and tourism support system,and ten specific image factors are recognized:hot tourism spots,urban topography,natural scenery,river and night view,ancient town,recreational facilities,special diet,history and culture,transportation,tourism services.River and night scenes in Chongqing are well-received tourisni hotspots.Natural landscape represented by the special geological structure of Wulong area has also left a deep impression on tourists.The evaluation of tourism services is bipolar,and tourists are dissatisfied with transportation,toilets and staff.In addition,the weather also has an important impact on travel experience.This research shows that natural language processing method based on word vector model can effectively draw perceived image elements,which facilitates the use of big data from social media to analyze image perception of urban tourism.This kind of method based on semantics further explores the information provided by user generated contents while reducing the manual element extraction and classification work,which makes the results more scientific and objective.