Predictive Analytics for Determining Learning Outcomes in Intelligent Tutoring Systems

Abstract

Educational structure in general and the learning process specifically are heading for a rapid transformation. Computers and computer networks are becoming major sources of information for the present and future generations. The emergence of computers and technology has opened up new avenues in the field of education. One of the major advancement is the development of Intelligent Tutoring Systems. This work introduces Intelligent Tutoring Systems along with their typical architecture, developmental history, past and presents systems and proposes a new concept of user profiling for better adaptivity in Intelligent Tutoring Systems. Cognitive factors such as concentration level, motivation level, mental speed, initial knowledge level and importance of task to the learner have been considered to have a huge impact on learning outcomes. Based on the values of these parameters, Naïve Bayes classifier has been used to classify the learner into three classes Beginner, Medium and Advanced. This classification is beneficial to determine the initial learning scenario and to provide personalization to each learner as every learner has its own cognitive skills and learning preferences. The amalgamation of emotions, human psychology, cognitive abilities, meta-cognitive abilities and the learning abilities with the ITS has introduced a new generation of ITS that is “Affective Tutoring Systems”.