합성곱 신경망과 이미지 피처 벡터 클러스터링을 활용한 국내 어촌마을 선호장면 분석

Abstract

Local extinction is now a reality due to the designation of half of South Korea's counties and cities as extinction risk areas. The risk stage of population extinction is predicted to reach all local governments in the nation by 2047, so local governments are actively using the tourism sector to combat local extinction. This study used image feature vector clustering and convolution neural networks to analyse the preferred scene of a domestic fishing village. After crawling the domestic fishing village data made available by the Instagram hashtag (#) using Python 3.9.7, data analysis was carried out. The analysis led to the classification of Waemok Village into 4 clusters, White Yeoul Culture Village into 7, Banwol-Bakji Village into 11, and Abai Village into 4 clusters. This allowed for the understanding of the primary tourist attractions in each fishing village, as well as the determination of the kinds of travel-related information that visitors share on social media. According to this study, the convolution neural network is a complementary methodology that offers an alternative to existing content analysis in terms of the use of large amounts of data. It can be used as a source of fundamental information for marketing plans and effective decision-making in related fields like local government, travel, and the planning and development of tourism products.