Predicting Learning Characteristics in a Multiple Intelligence Based Tutoring System

Abstract

Research on learning has shown that students learn differently and that they process knowledge in various ways. EDUCE is an Intelligent Tutoring System for which a set of learning resources has been developed using the principles of Multiple Intelligences. It can dynamically identify user learning characteristics and adaptively provide a customised learning material tailored to the learner. This paper introduces the predictive engine used within EDUCE. It describes the input representation model and the learning mechanism employed. The input representation model consists of input features that describe how different resources were used and inferred from fine-grained information collected during student computer interactions. The predictive engine employs the Naive Bayes classifier and operates online using no prior information. Using data from a previous experimental study, a comparison was made between the performance of the predictive engine and the actual behaviour of a group of students using the learning material without any guidance from EDUCE. Results indicate correlation between student’s behaviour and the predictions made by EDUCE. These results suggest that the concept of learning characteristics can be modelled using a learning scheme with appropriately chosen attributes.