Research On Tourism Destination Image Based On Sentiment Analysis And Theme Minin

In recent years,with the improvement of the quality of life,tourism has become a leisure way that people often choose.Based on this demand,online tourism service platforms such as tourism websites and tourism clients came into being.While providing tourism related services,they have gradually become an exchange platform for tourists to publish their evaluation of tourist attractions and understand the situation of tourist attractions.People with tourism plans will spontaneously search for tourism destination information through websites and clients.The most important one is to browse online texts such as tourist comments and travel notes,outline the image of tourism destination in their mind through others’ narration,judge the advantages and disadvantages,and then make decisions.It can be seen that the tourism destination image is no longer only based on the tourism resources and tourism activities of the tourism destination itself.The content of the tourism destination descriptive text published by tourists also has a vital impact on the formation of the tourism destination image.In order to construct the tourism destination image more comprehensively,this paper takes the online text published by tourists as the research object,and on the basis of reading a large number of literature,adopts the method of combining quantitative and qualitative methods to conduct a deeper research on the construction of tourism destination terrain image.Firstly,this paper expounds the current research situation of tourism destination image at home and abroad through literature review,and introduces the research and application of theme mining and emotion analysis.On the premise of understanding the relevant research status,this paper expounds the principles of word vector processing,cyclic neural network and machine learning classification algorithm,and defines the characteristics and applicable problems of different methods.Then,according to the research object and purpose of this paper,bilstm and LDA theme models are selected for modeling,and the analysis ideas of general steps,processes and results of constructing tourism destination image through the study of online texts published by tourists are determined.After completing the modeling,this paper takes Tianjin as the research object for empirical research.After crawling Ctrip Tianjin scenic spot comments and doing a series of preprocessing on the text data set,this paper begins to refine and analyze the tourist comments.Firstly,the text data is analyzed by bilstm,and the text data set is classified according to different emotions;Then the total text data set,positive emotion data set and negative emotion data set are respectively mined by LDA theme model,and then the mining results are visualized.Finally,according to the display results,the overall image of Tianjin as a tourism destination from the perspective of tourists,as well as the highlights and defects in the image are constructed.On this basis,combined with the classification and distribution of tourism resources in Tianjin,this paper gives reasonable suggestions to promote the development of tourism,so as to provide theoretical support for the development of tourism in Tianjin.Finally,this paper summarizes and prospects the overall research process,points out the areas that can be further improved and expanded in this research,and hopes that the research ideas of this paper can play a certain reference and inspiration for relevant research in the future.