메뉴 건너뛰기
소속 기관 / 학교 인증
인증하면 논문, 학술자료 등을  무료로 열람할 수 있어요.
한국대학교, 누리자동차, 시립도서관 등 나의 기관을 확인해보세요
(국내 대학 90% 이상 구독 중)
고객센터 ENG
주제분류

논문 기본 정보

저자정보
(Seoul National University of Science and Technology) (Seoul National University of Science and Technology)
저널정보
한국HCI학회 한국HCI학회 학술대회 PROCEEDINGS OF HCI KOREA 2023 학술대회 발표 논문집
오류 신고하기

검색

    초록·키워드

    As the use of online learning expands, it becomes increasingly important to predict and support at-risk students early. However, it is unclear which types of online learning behaviors (i.e., passive or active) are better able to predict at-risk students. Passive behavior (e.g., homepage access) refers to basic interaction with the system, while active behavior (e.g., foruming) refers to interactions involving other users. Using Open University Learning Analytics Dataset, we compared the predictive performance of passive and active behavior data to predict at-risk students. We used a random forest classifier to classify 3,994 students into either success or at-risk. Results showed that the predictive performance of passive behavior (accuracy: 0.78, precision: 0.79, recall: 0.91) was higher than that of active behavior (accuracy: 0.75, precision: 0.77, recall: 0.87) up to 20 weeks. These findings suggest the importance of fundamental passive behavior in online learning, such as accessing a homepage, compared to auxiliary active behavior. Through passive behavior analysis, instructors can predict at-risk students and help them successfully complete online courses.

    본문·목차

    최근 본 자료 전체보기