Analysis Method Of Tourist Destination Image Extraction On Integrating Microblog And GPS Trajectory

The widespread application of social software has brought a large amount of tourist UGC(User Generated Contents)data with rich content,which has become important data source for research on tourist destination image,tourist satisfaction,sentiment analysis,etc.However,the current research methods for these data are mostly limited to manual processing and analysis,and ignore the analysis of the spatio-temporal construction of tourist destination image.The paper proposes an image extraction method that integrates microblog and GPS(Global Positioning System)trajectory: firstly,based on the process of image construction,design time slicing methods to slice the data;secondly,use the LDA(Latent Dirichlet Allocation)model integrated feature dimensions to obtain the representative topics,and analyze the co-occurrence semantic network of the microblog text content of the representative topics at different time periods,and extract the main image with the central feature words,the key feature words of the co-occurrence semantic network are classified feature dimensions;lastly,for the representative topics with weak spatial features,the classified key feature words of the co-occurrence semantic network are used to obtain the process of time-based image construction,while for the topics with obvious spatial features,the corresponding GPS trajectory data are selected and the spatial and landscape feature words of the topic are coupled with the kernel density analysis results of the trajectory points and the trajectory photo points,and the behavioral and psychological feature words are coupled with the the rasterization results of the travel state vector in the trajectory points to obtain the process of spatio-temporal based image construction.The empirical data of Mount Tai tourism shows that the method of integrating microblogand trajectory data can refine the tourism destination image in terms of space,time and content,and the main conclusions are listed below.(1)The quantitative results of tourism destination image features are as follows: the quantitative index information of the first-level feature dimension of microblog LDA topic summarizes the core tourism image of Mount Tai: "climbing Mount Tai to watch the sunrise,experiencing a glimpse of the mountains;challenging the limits,surpassing oneself,having a heart for the world";the representative second-level dimension finely portrays the cultural heritage of Mount Tai as represented by "Wangyue",and the behavioral image of tourists climbing the mountain by various ways and enjoying the sunrise and clouds.Based on this,night climbing,cable car climbing,sunrise viewing and the Fengshan Ceremony viewing are selected as representative topics for detailed analysis.The reason why selecting the topic of Fengshan Ceremony instead of the topic of "Wangyue" for image analysis is that the topic of "Wangyue" includes mostly emotional text,the process of image evolution is not obvious,while the culture of Fengshan is a unique culture of Mount Tai,which is more representative and the process of viewing the performance is more valuable for analysis.(2)The construction of tourism destination image features are as follows:(1)the spatial characteristics of the topics of sunrise viewing and Fengshan ceremony viewing are not strong,so their temporal characteristics are analyzed: firstly,under the topic of sunrise viewing,the climing process for the purpose of sunrise viewing and the anxious state of waiting for sunrise is clearly shown by time slicing;secondly,under the topic of Fengshan ceremony viewing,the tourists’ understanding of the history of Mount Tai is gradually clarified,from the praise of the performance to the gradual understanding of the culture of Fengshan,and the change of the emotional state of praying for the family and the motherland is clearly shown;(2)the spatial characteristics of the night climbing and cable car climbing image are stronger,and the corresponding GPS trajectory data are integrated to analyze and obtain the construction of the spatial location and emotional characteristics of the tourists in the process of climbing: firstly,under the topic of night climbing,tourists climb fast at the beginning slow down in the middle;view sunrise after reaching the summit;enjoy the scenery leisurely during the descent;secondly,under the topic of cable car climbing,tourists climb easily by cable car,praise the scenery after reaching the summit,view sunrise with excitement,and climb hard during the climbing period without cable car.The results of integrating the two kinds of data provide a spatio-temporal construction analysis of the simple climbing image and obtain image results with richer spatio-temporal information,especially the cable car climbing imagery,where the microblog content mostly focuses on the ease of the cable car ride,but even when climbing by cable car,there is also a climbing process,which is highlighted by the integrated GPS trajectory data.Overall,the proposed method of extracting tourist destination image by integrating microblog and GPS trajectory provides a new way of thinking for the study of tourist destination image.The case study of Mount Tai demonstrates that the method can obtain richer spatio-temporal construction characteristics of image compared with single data souce.