日時	2024年7月10日(水) 16:45 – 17:45
場所	IMIオーディトリアム及びZoomによるオンラインコロキウム
講師	GIM, Minjung氏        National Institute for Mathematical Sciences/Ajou University（Korea） KWON, Soon-Sun 氏          Ajou University（Korea）
講演タイトル	GIM Minjung氏： Statistical Distances with Mathematical Explanation Soon-Sun Kwon氏：  Statistical Learning Models in Functional structure of Clinical data
講演要旨	Prof. GIM, Minjung (National Institute for Mathematical Sciences/Ajou University) 　      　”Statistical Distances with Mathematical Explanation” Statistical distances measure the difference between distributions or data samples and are employed in various  machine learning applications. In this talk, I will introduce several statistical distances and review their mathematical interpretations. We will demonstrate how to use SciPy’s statistical distance functions.  Using visual illustrations, we will describe the inner workings and properties of several common statistical distances, explaining what makes them both convenient to use and powerful for solving various problems. Additionally, we will present real-life applications and concrete examples.                Prof. KWON, Soon-Sun (Ajou University) “Statistical Learning Models in Functional structure of Clinical data” In this talk, I introduce two topics about longitudinal data analysis and gait data analysis. First, longitudinal data are used in statistical studies that accept many repeated measurements as well as  the different time spans of the measurements between or within subjects. Furthermore, correct inferences  can particularly be obtained by considering the correlation between repeated measurements within subjects.  Under the assumption, I propose the clustering method using the Fr’echet distance for multi-dimensional  functional data. And I apply the sparse clustering method to multi-dimensional thyroid cancer data collected in South Korea. Second, motivated by gait data from both the normal and the cerebral palsy (CP) patients  group with various gross motor function classification system (GMFCS) levels, I propose a multivariate functional classification method to investigate the relationship between kinematic gait measures and GMFCS levels. The method is generalized to handle multivariate functional data and multi-class classification. The method yields superior prediction accuracy and provides easily interpretable discriminant functions.
使用言語	英語

・zoomでの参加者は、事前のzoomのアプリ導入の推奨 (ダウンロード
サイト：https://zoom.us)

https://us06web.zoom.us/j/88249290750?pwd=DylHPXorzaRnNnbGaKQavu71I4wSva.1

ミーティングID: 882 4929 0750　　パスコード： 314361

・当日は、「教員はフルネーム」「学生は学籍番号(苗字)」(例：1SC○○○○○A(苗字))で
参加じてください
・聴講者の音声とビデオはホスト側でミュートに設定いたします
・質問したい際には、Zoomの挙手機能か、チャット機能でお知らせください
・当日は30分前にはログインできるように開催します

事前の質問を含め，もし何かございましたら田上先生
（E-mail：tagami@imi.kyushu-u.ac.jp）までご連絡ください。

※次回Colloquiumは10月9日（水）に開催予定です。

　　　　　

---

ニュース一覧に戻る イベントカレンダーに戻る