Research On Tourist Satisfaction Based On Web Text Mining

With the improvement of the living standards of Chinese citizens gradually,tourism has become an increasingly popular topic.Inner Mongolia is rich in tourism resources and consistently ranks at the top of the tourism topic list.More and more tourists choose to order service products on tourism service platforms,resulting in a large amount of user generated content.Studying online reviews can identify factors that affect tourist satisfaction,which is of great significance in improving tourist satisfaction and promoting the development of scenic spots.It also provides more decision-making information for the rapid recovery of vitality in Inner Mongolia scenic spots and the prediction of tourist demand in the early stages of the post pandemic.This article uses text mining to identify the factors that affect the satisfaction of tourists in Inner Mongolia scenic areas,and constructs a fuzzy comprehensive evaluation index system using the Analytic Hierarchy Process(AHP)to calculate the satisfaction of each indicator layer.Finally,combined with the Baidu Index,it predicts the satisfaction of tourists in the post epidemic period from a spatiotemporal perspective.Firstly,this article crawled a total of 27385 Inner Mongolia tourism online comment text data on the Ctrip platform using a Python program.After data cleaning,text segmentation,and other operations,the text preprocessing of the corpus was completed,resulting in 19862 valid comments.The basic analysis of the text comment data was conducted to preliminarily understand the characteristics of the data.Then,through TF-IDF and TextRank high-frequency word analysis,we can understand tourists’ attention to various aspects of tourism in Inner Mongolia,and use syntactic dependency tree to analyze the relationship between words,use semantic network to analyze the semantics of sentence segments,and use ROSTCM6 software to visually analyze the internal relationship between the influencing factors of semantic network.Then,the LDA topic model is used to mine the subject words of the evaluation index system.After determining the optimal number of topics in the review text corpus using visual methods,the topics and corresponding feature words of the Inner Mongolia tourism review text are obtained.It is found that the Inner Mongolia tourism online review mainly includes eight topics:scenic spots,accommodation and transportation,scenic spot consumption,tourism services,travel experience,willingness to come again,epidemic prevention and control,and environmental atmosphere.Secondly,by reviewing literature and combining with text mining results,a tourist satisfaction index system for Inner Mongolia scenic spots was constructed,which includes five dimensions.Using MATLAB software,a fuzzy comprehensive evaluation index system based on AHP was constructed to evaluate tourist satisfaction.The results showed that the five dimensions that affect tourist satisfaction in Inner Mongolia scenic spots were ranked according to their importance:scenic spots,sightseeing,service quality Tourism experience and impact of the epidemic.The overall evaluation score of tourist satisfaction in Inner Mongolia scenic areas is 4.472,which is between excellent and satisfactory.Overall,tourists’ satisfaction evaluation of Inner Mongolia is relatively high.Finally,the IPA analysis method was used to construct a two-dimensional four quadrant grid of "importance satisfaction",and it was found that the urgently needed areas for improvement in the scenic area include scenic spot consumption,gaming services,and viewing experience.In order to quickly restore the vitality of Inner Mongolia scenic spots in the early stage of the epidemic,this article analyzes the attention of tourists to Inner Mongolia scenic spots from a spatiotemporal perspective based on the Baidu index.It is found that there is significant seasonal fluctuation in tourist tourism in Inner Mongolia scenic spots,and there is a clear geographical neutrality in tourist source areas.Therefore,a multivariate LSTM model is used to predict tourist satisfaction with the scenic spot.The overall MSE of the model is 0.22,and the model performs well.The prediction results show that:,The overall satisfaction of tourists from April to June was 4.648,showing a slight upward trend compared to the satisfaction of tourists during the epidemic period.