| New data sources and perspectives have been provided for the extraction of tourism destination image thanks to plenty of user generated contributions on social media.In such a context,this paper designed an improved topic-dimension-relationship framework against the dependency on researcher experience and the vague recognition criteria in the topic-dimension framework during extracting tourism destination images.(1)The basic theoretical basis and criteria for the determination of the number of topic clusters are given in the design of the UGC text word separation optimization method based on feature thesaurus and deactivation word list in the topic clustering session.Clarify the recursive process of refining the concept of topic from a purely mathematical model to the semantics of tourist place image,clarify the concept of dimension is the division of semantic properties or categories of feature words,and is the information representation of the research object or content from a goal-oriented multidimensional structure.We propose a revised idea based on the existing dimensional division scheme,design a new two-level dimensional division standard,and introduce a method to reduce the error of dimensional division,and control the direction of each stage of the image extraction process of tourist places through dimensional division.(2)On this basis,a relationship module was added for recognizing common logic implied in the construction of tourism destination image,so as to define multi-type and multi-level relationship analytical method required for extracting theme dimension analysis and image extraction.Designing methods for sequential,structural and co-occurrence relationships,and pointing out the way,criteria and significance of the application of various types of relationships at various levels of detail in the process of image feature extraction,for identifying the common logic implicit in the construction of image in tourist places,and identifying core image elements,structural frameworks or potential clues to image features through relationship analysis.(3)Based on the conceptual framework,a detailed technical implementation plan and specific application methods including 4 main links are designed.By doing so,the extraction of semi-quantitative interactive tourism destination image can be achieved from quantitative analysis constraints and key clue verification.The image of tourist places is portrayed at three levels: overall image,fine image,and perceptual experience characteristics,while the spatio-temporal distribution patterns of different themes are explored to provide ideas for the spatialization of image.Based on the new framework,two image characteristics of the ancient town tourism site are refined: the first level dimension provides a macro overview of the overall image characteristics of the ancient town.Also,in the second level dimension,its multi-level exquisite characteristics are portrayed with core image symbols,space-time axes,primary areas,key nodes as well as space and cultural locations recognized.Furthermore,the author also stimulated the future and reflections on life beyond the meaning of travel upon disclosing the processes of various spatial images from nature,history,culture,and literature integrating into complex image processes.Based on this,we summarize and condense the image characteristics of ancient towns and their commonalities and differences.In that case,the proposed framework contributes to offering a new technical scheme for the extraction and meticulous depict of tourism destination images,the spatial and temporal distribution and the comparative analysis of the image.More than that,it is also applicable for the delicacy management and building of tourism destination images upon clarifying the image forming process and driving factors. |
PDF Full Text Request