Research On Hunan Tourist Attractions Recommendation Based On Subspace Clustering

Along with the rapid development and rise of the "Internet+Tourism" industry model,tourists are eager to publish travel information in the form of travelogues on online platforms,and the application and development of travel recommendation methods and geographic information systems have made it possible to explore the recommended places in travelogues and explore their spatial distribution characteristics.This study is based on 3451 online travelogues about Hunan region available on the Hornet’s Nest website from January 1,2018 to June 9,2022.First,by pre-processing the travelogue data in Python,a text dataset suitable for building the clustering matrix was derived,keyword extraction and Chinese word separation were performed on the travelogue data,and then a travelogue author-location matrix was constructed,the matrix data was imported into the subspace clustering algorithm,and the clustering results were evaluated and compared using the contour coefficient and CH index evaluation index,and the recommended sets of national 4A and 5A scenic spots and district and county cities in Hunan Province under the optimal clustering were derived;Secondly,we counted and obtained the word frequencies and geographic coordinates of the recommended concentrated scenic spots and cities,and analyzed the spatial distribution characteristics of the recommended scenic spots and cities respectively by using the tools of nearest neighbor index,standard deviation ellipse,and kernel density analysis in ArcGIS10.2.The specific conclusions are as follows:(1)Based on the subspace clustering algorithm,the clustering analysis and clustering evaluation of the "Author-Scenic spot" matrix and the "Author-City" matrix are performed,and the optimal clustering effect is obtained after adjusting the parameters.The word frequency statistics and coordinate picking of the output attraction-city set recommendations yielded word frequency rankings and geographic coordinates of attractions and cities,and found that the western Xiang region,represented by Zhangjiajie and Huaihua,and the eastern Xiang region,represented by Changsha,were the tourist destinations recommended by tourists with higher frequency.(2)The national 4A and 5A recommended scenic spots in Hunan Province are clustered,showing the overall characteristic of "dense in the east and sparse in the west",distributed along the northwest-southeast direction,with a more obvious centripetal nature.From the type of spatial distribution,the frequency of recommended scenic spots in Changsha is higher than other cities,and scenic spots such as Orange Island,Yuelu Mountain,Tianxin Pavilion,Jinggang Ancient Town,Window of the World and Furong Ancient Town are more attractive to tourists,and tourists coming to Hunan are more likely to visit scenic spots in Changsha;Secondary area for Zhangjiajie and other western cities and Zhuzhou,Yongzhou and other southeastern areas to natural scenic spots,is the second largest choice for tourists to travel to Hunan after Changsha,is a better choice for tourists to come to Hunan to enjoy the natural scenery;The third-level areas are the northern cities such as Changde and central cities such as Loudi and Yiyang and Shaoyang in the fourth-level areas.The number of high-level scenic spots in such areas is small,the popularity is low,the recognition of tourists is relatively weak,and the construction of tourism infrastructure is not perfect compared with Changsha and Zhangjiajie,resulting in less frequent recommendation by tourists.From the viewpoint of spatial clustering characteristics,the overall characteristics of"east is dense and west is sparse",the city cluster of ChangZhuTan in the eastern part of Hunan Province applies and constructs high-level scenic spots more than the western cities,and more tourists come here than in the western part of Hunan Province;the scenic spots recommended by tourists are distributed along the northwest-southeast direction,from ChangZhuTan area to Zhangjiajie and other areas in the western part of Hunan Province,with the center point near Shaoshan scenic spot in Xiangtan,and the centripetal nature is more obvious.(3)The recommended tourism cities in Hunan Province are scattered distribution characteristics,with no obvious clustering characteristics,northwest-southeast direction,and weak centripetal.From the spatial distribution,influenced by the regional scope and natural environment,cities in western Hunan such as Huaihua,Zhangjiajie and Xiangxi Autonomous Prefecture are recommended with high frequency,and various district and county-level cities are influenced by them in tourism to a certain extent,which is also related to the rich tourism resources around most cities in the region.Comprehensive analysis of the existing location conditions and geographical environment in the west and east,the recommended frequency of ChangZhuTan and the surrounding cities is lower than that of the western Hunan region,indicating that the tourism development in ChangZhuTan region is dominated by the central cities,and the tourism market space of the district and county-level cities within its jurisdiction is insufficient,and the recommended frequency of travel to such cities is low.From the spatial distribution agglomeration situation,the recommended cities are generally northwest-southeast direction,mainly cities in the southeast and northwest,while Loudi,Shaoyang and other cities in the central and southwestern regions due to the lower economic level and low visibility,resulting in lower tourist recommendation and recognition.The center point of the recommended cities is in the central city of Loudi,and the centripetal nature is not obvious.Finally,based on the above conclusions,development suggestions such as upgrading technology to improve the quality of tourism recommendations,living resources to enhance the competitiveness of tourism regions,and building tourism city brands according to local conditions are proposed.