A conception for use of user profile to prediction learning effects in Intelligent Tutoring Systems

Abstract

Intelligent Tutoring Systems (ITS) offer adaptivity to user abilities, personal character trait, learning styles and preferences. The user modelling is one of the major factors that can influence an adaptivity. The content and structure of user profile should allow to recommend learning material suitable for student's needs. In this work a user profile is designed and a method for predicting learner's abilities is proposed. We use a Naive Bayes classifier in order to predict student's learning results. A prediction of user's abilities could be very useful for determining an initial learning scenario or assigning a student to a suitable collaborative learning group.