Semi-Supervised Machine Learning for Domain Modeling in Network-Based Tutoring Systems: Implications for Fostering Self-Regulated Learning

• In Event: 61.069 - MTCC Roundtable Session 24
  In Roundtable Session: 61.069-9 - Data Analytics and Research Methodologies in Cognition and Learning

Abstract

Automated methods to create a computational representation of a domain is a major challenge for developers of adaptive open-ended learning environments. Too often, the ability of a system to adapt instruction is limited to the system developers’ capability to capture a broad range of skills, knowledge, and strategies. In this study, we demonstrate the use of a semi-automated machine learning approach to discover latent dimensions in textual documents. We apply this approach to model content extracted from open educational resources as a means to support teacher professional learning with nBrowser, a network-based tutor designed for pre-service teachers to facilitate lesson planning and technological integration. We discuss the broader implications for the design of recommender systems in open-ended learning environments.

Authors

• Eric G. Poitras, University of Utah

• Laurel Udy, University of Utah

• Lingyun Huang, McGill University

• Shan Li, McGill University

• Susanne P. Lajoie, McGill University