@techreport{oai:ipsj.ixsq.nii.ac.jp:00233846, author = {呉, 祖煕 and 単, 詩軒 and 中村, 亮太 and Soki, Go and Shiken, Zen and Ryota, Nakamura}, issue = {7}, month = {May}, note = {近年,オンラインコミュニティの普及に伴い,バーチャル空間におけるユーザ体験の向上が求められている.本研究では,バーチャル空間で初対面のユーザを対象に,アイスブレイク中の発言状況を使い,相性の良いグループを形成できるよう,バーチャル空間での発言データを機械学習で分析し,ユーザの特性や適性に応じたグループ分類手法を提案している.実証実験の結果,従来手法よりも的確にグループ内の発言が均等化され,議論が効率化されることが示された.提案手法はオンラインコミュニティの活性化と参加者満足度の向上に貢献できる., In recent years, with the widespread adoption of online communities, there has been an increasing emphasis on enhancing user experiences in virtual environments. This study aims to address this demand by focusing on improving interactions among users meeting for the first time in virtual spaces. Utilizing speech data during breakout sessions, we propose a machine learning-based approach to analyze speech patterns and form compatible groups tailored to users' characteristics and aptitudes. Our empirical experiments demonstrate that our proposed method achieves more balanced participation within groups compared to traditional methods, leading to enhanced efficiency in discussions. Overall, our approach contributes to the revitalization of online communities and improves participant satisfaction.}, title = {オンラインコミュニティーにおけるユーザ属性に基づくグループ分けシステムの試作と評価}, year = {2024} }