Identifying Learning Styles from Chat Conversation using Ontology-Based Dynamic Bayesian Network Model

One of the commonly adopted parameters in personalized e-learning is learning style. As an individual trait, this parameter indicates the preferable learning object for a specific type of a learner. When considering collaborative learning environment through online chat and discussion, the learners can express their opinion on a shared learning object. Besides, level of interaction with the object is also important to identify the learning style. The nature of generated contents during discussion makes it difficult to extract the required information. In this paper we discuss how to identify the learning style from the learning object preferences expressed by the learner via the online discussion in a collaborative learning platform. To do so, the paper proposes an ontology-based Dynamic Bayesian Network (DBN) model to represent the relationship between the learning style and preferable learning object. The model also obtains the learner's opinion more than one time by using time slice to make the indication of learning styles more accurate. Consequently, providing the learner the appropriate personalized learning package.