Java Error Correction Algorithm in the Framework of an Intelligent Tutoring System

Abstract

<p>Intelligent Tutoring Systems (ITS) are, in many respects, very similar to human tutors. Based on cognitive science and Artificial Intelligence (AI), ITS have proven their significance in many disciplines. Currently, ITS can be found in core Mathematics, Physics, and Language courses in hundreds of schools across Canada, the United States, and various countries in Europe. ITS are growing in acceptance and popularity for reasons including: i) increased student performance, ii) deepened cognitive development, and iii) reduced time for student to acquire skills and knowledge.</p> <p>Bloom (1984) showed that one-on-one human tutors could increase the average student's performance to the ninety-eighth percentile in a standard classroom. Furthermore, in order for students to reach their potential, individualized tutoring is a necessity. Intelligent Tutoring Systems have demonstrated that student achievement is 1.0 standard deviation higher than typical classroom environments. Another benefit ITS have is the speed of knowledge acquisition. Students learning from an ITS have completed problems in one-third of the time compared to students in the control group.</p> <p>This thesis focuses on the design and implementation of an error correction algorithm in the specific context for use in an ITS for the Java programming language. The Java Error Correction Algorithm (JECA) was designed to be used by first year College and University students with little or no programming experience. JECA attempts to determine the "intent" of the student's submission by rigorously analyzing the student's code. Behind the scenes, JECA makes changes to the student's submission in order to facilitate this analysis. However, once JECA determines the most reasonable intent of the student, these changes are made known to the student. The results from JECA are passed to the Java Intelligent Tutoring System (JITS) in the form of hints and suggestions, which are then used for instructional purposes.</p> <p>This thesis focuses on JECA, however, to ensure contextual relevance and significance, the Java Intelligent Tutoring System is included. JITS is implemented using advanced e-learning technologies and its multi-threaded distributed architecture makes JITS scalable, robust and easy to maintain. JITS supports personalized student development by tracking and modeling every student in the system. JITS is an online website always available for students and requires only a browser and an internet connection.</p>