Home // ICIW 2020, The Fifteenth International Conference on Internet and Web Applications and Services // View article

Improving DMF with Hybrid Loss Function and Applying CF-NADE to the MOOC Recommendation System

Authors:
Thanh Le
Vinh Vo
Khai Nguyen
Bac Le

Keywords: MOOCs; Recommendation; CF-NADE; DMF

Abstract:
Nowadays, with the strong development of platforms like Coursera, Edx, etc., Massive Open Online Course (MOOC) is not too strange for most people. The number of online courses also increases day by day. One of the problems raised is how to recommend users to choose the appropriate course. To address the problem, we applied the Deep Matrix Factorization (DMF) model to build a user-item interaction matrix with explicit rating and zero implicit feedback. We then improved the loss function to yield more accurate results. In addition, we also used the Collaborative Filtering Neural Autoregressive Distribution Estimator (CF-NADE) model to MOOC Recommendation system. Our experiment shows that two proposed approaches achieve better results than the other methods.

Pages: 13 to 20

Copyright: Copyright (c) IARIA, 2020

Publication date: September 27, 2020

Published in: conference

ISSN: 2308-3972

ISBN: 978-1-61208-803-7

Location: Lisbon, Portugal

Dates: from September 27, 2020 to October 1, 2020