Architecture of an Adaptive Personalized Learning Environment (APLE) for Content Recommendation

With the development of sophisticated learning environments and learner-centric didactic approaches, personalized learning is in high demand. Personalization in learning environments occurs when such systems fit the learner profiles, which help in increasing their performance and quality of learning. Personalized learning refers to the pedagogy where the pace of learning, the instructional preferences and the learning objects are optimized as per the needs of each learner. To support customization, recommender systems can be used to recommend appropriate learning objects (LOs) corresponding to the learner attributes. This paper proposes an architecture of an Adaptive Personalized Learning Environment (APLE) and its features. APLE assists the learners by content recommendation and adapts to the learning preferences and performance of the learner. It has three modules such as Learner modelling Unit (LModU), Content Managing Unit (CMU) and Learner Monitoring Unit (LMU). LModU creates a Learner Model (LM) from the learner attributes. The system proposes to represent learner attributes as an ontology and learner modelling using Dynamic Bayesian Networks. The LMU should perform the knowledge assessment of the learner and monitor their changing preferences. CMU got two components, LO Manager and Content Recommendation Engine (CRE). LO Manager is responsible for creating the metadata corresponding to the learning resource following the IEEE LOM specification. CRE is an expert system which will map the learner attribute with the LOs. Currently, the CRE is implemented as a rule-based prediction engine where the rules represent the association between each LOM field with the learner attributes. This on-going research work aims at answering questions regarding the feasibility and effectiveness of mapping LO attributes and learner attributes.