Thread Structure Prediction for MOOC Discussion Forum

摘要

Discussion forums are an indispensable interactive component for Massive Open Online Courses(MOOC).However,the organization of current discussion forums is not well-designed.Trouble-shooting threads are valuable for both learners and instructors,but they are drowned out in the forums with huge amounts of threads.This work first built a labeled data set for trouble-shooting thread structure prediction by crowdsourcing and then proposed methods for trouble-shooting thread detection and thread structure prediction on the data set.The output of this work can be used to spot trouble-shooting threads and show them along with structure tags in MOOC discussion forums.