作者: Min-Yen Kan , Muthu Kumar Chandrasekaran , Kiruthika Ragupathi , Bernard C. Y. Tan
DOI:
关键词:
摘要: With large student enrollment, MOOC instructors face the unique challenge in deciding when to intervene forum discussions with their limited bandwidth. We study this problem of instructor intervention. Using a sample data culled from 61 courses, we design binary classifier predict whether an should discussion thread or not. By incorporating novel information about forum's type into classification process, improve significantly over previous state-of-the-art. We show how difficult decision is real world by validating against indicative human judgment, and empirically problem's sensitivity instructors' intervention preferences. conclude paper our take on future research issues