An adaptive framework for recommender-based Learning Management Systems

There are a number of existing frameworks for recommendation systems that have been identified in domains such as e-commerce and tourism. The aspects of user profiles, adaptation and group models from the e-commerce and tourism frameworks can be applied to education provided they are customised through incorporating principles of pedagogy such as behavioural theories. Recommendation systems can be adopted to support personalised / differentiated teaching and learning. The aim of this research is to develop an adaptive recommender-systems-based framework for differentiated teaching and learning on eLearning platforms, that is, learning management systems (LMS). Through a literature review, 40 attributes of personalized learning were identified. The Multi-Attribute Utility Theory (MUAT) was used to identify the 10 top attributes to go in as personalized learning framework components. From a population of 1203 students from College X, a sample of 200 students was purposively selected for the research on the basis of their familiarity with College X's eLearning system. 103 students responded to the questionnaire, representing a response rate of 52%. From the responses of the students, the following top ten (10) attributes were identified for inclusion in the personalised learning platforms: culture, emotional/mental state, socialisation, motivation, learning preferences, prior knowledge, educational background, learning/cognitive style, and navigation and learning goals. A theory-driven adaptive recommender-based framework was derived from a combination of literature review and the attributes derived from the research.