ABSTRACT
In a Massive Open Online Course (MOOC), predictive models of student behavior can support multiple aspects of learning, including instructor feedback and timely intervention. Ongoing courses, when the student outcomes are yet unknown, must rely on models trained from the historical data of previously offered courses. It is possible to transfer models, but they often have poor prediction performance. One reason is features that inadequately represent predictive attributes common to both courses. We present an automated transductive transfer learning approach that addresses this issue. It relies on problem-agnostic, temporal organization of the MOOC clickstream data, where, for each student, for multiple courses, a set of specific MOOC event types is expressed for each time unit. It consists of two alternative transfer methods based on representation learning with auto-encoders: a passive approach using transductive principal component analysis and an active approach that uses a correlation alignment loss term. With these methods, we investigate the transferability of dropout prediction across similar and dissimilar MOOCs and compare with known methods. Results show improved model transferability and suggest that the methods are capable of automatically learning a feature representation that expresses common predictive characteristics of MOOCs.
- Andrew Arnold, Ramesh Nallapati, and William W Cohen. 2007. A Comparative Study of Methods for Transductive Transfer Learning. In ICDM Workshops. ICDM, Pittsburgh, PA, USA, 77--82. Google ScholarDigital Library
- Ryan Shaun Baker and Paul Salvador Inventado. 2014. Educational data mining and learning analytics. In Learning analytics. Springer, New York, NY, USA, 61--75.Google Scholar
- Shai Ben-David, John Blitzer, Koby Crammer, Alex Kulesza, Fernando Pereira, and Jennifer Wortman Vaughan. 2010. A theory of learning from different domains. Machine learning 79, 1-2 (2010), 151--175. Google ScholarDigital Library
- Sebastien Boyer and Kalyan Veeramachaneni. 2015. Transfer Learning for Predictive Models in Massive Open Online Courses. In Artificial Intelligence in Education, Cristina Conati, Neil Heffernan, Antonija Mitrovic, and M. Felisa Verdejo (Eds.). Springer International Publishing, Cambridge, MA, USA, 54--63.Google Scholar
- Christopher Brooks, Craig Thompson, and Stephanie Teasley. 2015. A time series interaction analysis method for building predictive models of learners using log data. In Proceedings of the fifth international conference on learning analytics and knowledge. ACM, 126--135. Google ScholarDigital Library
- Concepción Burgos, María L. Campanario, David de la Peña, Juan A. Lara, David Lizcano, and María A. Martínez. 2018. Data mining for modeling students' performance: A tutoring action plan to prevent academic dropout. Computers & Electrical Engineering 66 (2018), 541--556. Google ScholarDigital Library
- Devendra Singh Chaplot, Eunhee Rhim, and Jihie Kim. 2015. Predicting Student Attrition in MOOCs using Sentiment Analysis and Neural Networks.. In AIED Workshops (CEUR Workshop Proceedings), Jesus Boticario and Kasia Muldner (Eds.), Vol. 1432. CEUR-WS.org, Seoul, South Korea.Google Scholar
- Josh Gardner and Christopher Brooks. 2018. Student success prediction in MOOCs. User Modeling and User-Adapted Interaction 28, 2 (2018), 127--203. Google ScholarDigital Library
- Josh Gardner, Christopher Brooks, Juan Miguel Andres, and Ryan Baker. 2018. Replicating MOOC Predictive Models at Scale. In Proceedings of the Fifth Annual ACM Conference on Learning at Scale (L@S '18). ACM, New York, NY, USA, 1:1--1:10. Google ScholarDigital Library
- Jiayuan Huang, Alexander J. Smola, Arthur Gretton, Karsten M. Borgwardt, and Bernhard Scholkopf. 2006. Correcting Sample Selection Bias by Unlabeled Data. In Proceedings of the 19th International Conference on Neural Information Processing Systems (NIPS'06). MIT Press, Cambridge, MA, USA, 601--608. Google ScholarDigital Library
- Xin J. Hunt, Ilknur Kaynar Kabul, and Jorge Silva. 2017. Transfer Learning for Education Data. KDD Workshop (2017).Google Scholar
- Ian T Jolliffe. 1986. Principal component analysis and factor analysis. In Principal component analysis. Springer, Kent, England, UK, 115--128.Google Scholar
- Marius Kloft, Felix Stiehler, Zhilin Zheng, and Niels Pinkwart. 2014. Predicting MOOC dropout over weeks using machine learning methods. In Proceedings of the EMNLP 2014 Workshop on Analysis of Large Scale Social Interaction in MOOCs. Association for Computational Linguistics, Doha, Qatar, 60--65.Google ScholarCross Ref
- Christopher Vu Le, Zachary A. Pardos, Samuel D. Meyer, and Rachel Thorp. 2018. Communication at Scale in a MOOC Using Predictive Engagement Analytics. In Artificial Intelligence in Education - 19th International Conference, AIED 2018, London, UK, June 27-30, 2018, Proceedings, Part I. IJAIED, Berkeley, CA, USA, 239--252.Google Scholar
- Grégoire Mesnil, Yann Dauphin, Xavier Glorot, Salah Rifai, Yoshua Bengio, Ian Goodfellow, Erick Lavoie, Xavier Muller, Guillaume Desjardins, David Warde-Farley, Pascal Vincent, Aaron Courville, and James Bergstra. 2012. Unsupervised and Transfer Learning Challenge: a Deep Learning Approach. In Proceedings of ICML Workshop on Unsupervised and Transfer Learning (Proceedings of Machine Learning Research), Isabelle Guyon, Gideon Dror, Vincent Lemaire, Graham Taylor, and Daniel Silver (Eds.), Vol. 27. PMLR, Bellevue, Washington, USA, 97--110. Google ScholarDigital Library
- Sinno Jialin Pan and Qiang Yang. 2010. A survey on transfer learning. IEEE Transactions on knowledge and data engineering 22, 10 (2010), 1345--1359. Google ScholarDigital Library
- Baochen Sun, Jiashi Feng, and Kate Saenko. 2016. Return of frustratingly easy domain adaptation. In AAAI, Vol. 6. Palo Alto, CA, USA, 8 pages. Google ScholarDigital Library
- Baochen Sun and Kate Saenko. 2016. Deep coral: Correlation alignment for deep domain adaptation. In European Conference on Computer Vision. Springer, Springer International Publishing, Cambridge, MA, USA, 443--450.Google ScholarCross Ref
- Kalyan Veeramachaneni, Una-May O'Reilly, and Colin Taylor. 2014. Towards Feature Engineering at Scale for Data from Massive Open Online Courses. CoRR abs/1407.5238 (2014).Google Scholar
Index Terms
- Transfer Learning using Representation Learning in Massive Open Online Courses
Recommendations
Effective Feature Learning with Unsupervised Learning for Improving the Predictive Models in Massive Open Online Courses
LAK19: Proceedings of the 9th International Conference on Learning Analytics & KnowledgeThe effectiveness of learning in massive open online courses (MOOCs) can be significantly enhanced by introducing personalized intervention schemes which rely on building predictive models of student learning behaviors such as some engagement or ...
Supporting learners' self-regulated learning in Massive Open Online Courses
AbstractIn MOOCs, learners are typically presented with great autonomy over their learning process. Therefore, learners should engage in self-regulated learning (SRL) in order to successfully study in a MOOC. Learners however often struggle to self-...
Highlights- Learners struggle to regulate their learning in massive open online courses (MOOCs).
- A self-regulated learning (SRL) intervention was implemented in three MOOCs.
- Learners' SRL was measured with trace data variables.
- ...
Benefit and Cost Analysis of Massive Open Online Courses: Pedagogical Implications on Higher Education
There has been much research done on online learning including research on online educational activities and methods. The use of technology is gaining rising importance in higher education due to the benefits that it brings. In terms of adopting new ...
Comments