Domain-based intelligent tutoring system

Abstract

In this paper we present a new framework for predicting the proper instructional strategy for a given teaching material based on its attributes. The framework is domain-based in the sense that it is based on the qualitative observation of the teaching materials' attributes stored by the system. The prediction process is based on a machine learning approach using feed forward artificial neural network to generate a model that both fit the input data attributes and predict the proper instructional strategy by extracting knowledge implicit in these attributes. The framework was adapted in an Intelligent Tutoring System (ITS) to teach Modern Standard Arabic language to adult English-speaking learners with no pre-knowledge of Arabic language is required. The learning process will be through the Internet since the online education is better suited to mature individuals who are self-motivated and have a good sense of purpose. © Springer Science+Business Media B.V. 2010.