文献类型：期刊文献

英文题名：A CNN-Bi-LSTM Model for MOOC Forum Post Classification

作者：Zhang, Qiaorong[1]; Sun, Lin[2]

第一作者：张巧荣

机构：[1] College of Computer and Information Engineering, Henan University of Economics and Law, Zhengzhou, China; [2] School of Economics and Management, China University of Petroleum, Beijing, China

第一机构：河南财经政法大学计算机与信息工程学院

年份：2023

卷号：18

期号：21

起止页码：89-101

外文期刊名：International Journal of Emerging Technologies in Learning

收录：EI(收录号：20234815110292)

语种：英文

外文关键词：Behavioral research - Brain - E-learning - Learning systems - Taxonomies - Teaching

摘要：The discussion forum of the massive open online course (MOOC) is a platform for students to communicate with teachers, teaching assistants, and platform managers. It is one of the important factors related to course quality. A reasonable classification of discussion posts in the forum will help students better communicate and solve problems, so as to improve the quality of teaching. Aiming at the classification of discussion forum posts, this paper proposes a text classification model integrating convolutional neural networks (CNN) and bidirectional long-short-term memory (Bi-LSTM). Firstly, the user types and behavior characteristics are analyzed to build the taxonomy. The taxonomy includes three categories: course related, teacher related and platform related. Then, a text classification model is constructed based on CNN and Bi-LSTM. In order to verify the effectiveness of the proposed model, it is applied to the classification of 19285 discussion posts from the MOOC platform of icourse163.org. The overall classification accuracy of the proposed model is 93.6%, which is 12%, 10%, and 8% higher than traditional machine learning methods, CNN and Bi-LSTM, respectively. The model is used for automatic text classification in MOOC discussion forum, which can provide effective help and support for learners, teachers and platform managers, and improve the automation level of MOOC platform. ? 2023 by the authors of this article. Published under CC-BY. All Rights Reserved.