A Comparative Study On Image Perception Of Hainan Island Tourism And Taiwan Island Tourism Based On Massive UGC Image Metadata

With the gradual development of online tourism platforms,the influence of User Generated Content(UGC)on tourists’ travel behavior has gradually increased.By analyzing the research background,theoretical basis and research methods of island tourism image perception,this paper adopts the methods of literature analysis,the combination of qualitative and quantitative methods,image analysis and comparative analysis,and uses crawler to capture the UGC pictures of Hainan tourism and Taiwan tourism generated by OTA platform users in the past five years.The deep convolutional neural network model(Deep Senti Bank)was used to analyze the crawled UGC images and generate ANP sequence,that is,the combination of "adjective + noun".Highfrequency word statistics and analysis were carried out on nouns and adjectives respectively,and tourists from different sources summarized the cognitive image and emotional image of Hainan and Taiwan tourism destinations.The findings are as follows: 1.The cognitive image of Hainan and Taiwan tourist destinations is mainly concentrated in the natural scenery,but the proportion of image attributes is different;2.Most domestic and foreign tourists have a positive emotional image of Hainan and Taiwan as tourist destinations,and Hainan and Taiwan are favored by tourists for vacation.Finally,based on the actual situation,the paper analyzes the reasons for the differences in tourism image perception between Hainan and Taiwan,and puts forward suggestions for promoting tourism and improving tourists’ perception in Hainan and Taiwan respectively.