@techreport{oai:ipsj.ixsq.nii.ac.jp:00208639, author = {陸, 昱宏 and 谷口, 雄太 and 木實, 新一 and Yuhong, Lu and Yuta, Taniguchi and Shin'ichi, Konomi}, issue = {25}, month = {Dec}, note = {少子高齢化の進行により,高齢者の孤独・孤立が深刻な問題になっている.こういった問題を抱える高齢者を特定して問題解決に導くための提案を行うことができれば,高齢者本人のみならず社会にもさまざまな効用をもたらす.本論文では,高齢者に観光を促すことにより,人との繋がりを作り,孤独や孤立を解消することを考える.SNS 上で高齢者ユーザーが発信する内容に示された興味に着目し,Semi-Supervised Latent Dirichlet Allocation モデルを画像に含まれる情報と統合し,生成した興味トピックと観光地の特徴を比較することで効果的に観光推薦を行う手法を提案する., Due to the declining birthrate and aging population, the loneliness and isolation of older adults have become a serious problem. If we can identify older adults with such problems and make proposals to solve them, it will bring various benefits not only to older adults themselves but also to the society. In this paper, we consider encouraging older adults to go sightseeing to create connections with people and eliminate loneliness and isolation. We focus on the interests shown in the content sent by older adults on SNS, propose a method to integrate the Semi-Supervised Latent Dirichlet Allocation model with the information contained in the images, and compare created interesting topics and tourist destination characteristics to recommend tourist spots effectively.}, title = {SNSにおけるテキストおよび画像として表現された潜在的興味に基づく高齢者向けの観光推薦手法の提案}, year = {2020} }