A Framework for Developing Intelligent Tutoring Systems Based on Domain Models in the Form of Decision Trees

Formal domain models make it possible to build intelligent tutoring systems based on symbolic artificial intelligence that are capable of reliable error detection, generation of explanatory feedback and pedagogical questions, etc. However, developing these models for modern inference engines is a very time-consuming problem. This paper describes a method for representing domain problems and solutions in the form of decision trees and proposes a prototype framework for developing intelligent tutoring systems based on it. During the experimental development of four kernels for intelligent tutoring systems, the framework showed moderate performance, providing a significant reduction in development time.