Journal for Research | Volume 03| Issue 10 | December 2017

ISSN: 2395-7549

Comparative Analysis of Intelligent Tutoring

System Approaches
Pankaj Yadav
Department of Computer Science and Engineering
Thapar Institute of Engineering and Technology
(Deemed to be University), Patiala, India
Harkiran Kaur
Department of Computer Science and Engineering
Thapar Institute of Engineering and Technology
(Deemed to be University), Patiala, India

Abstract
With the huge advancement in technology there is a need of advancement in education. Lots of work have been done in
education system to enhance the learning ability of students, Intelligent Tutoring System (ITS) is one of the outcome of those
advancements. It provides a platform which reduce the dependency on human tutor as it focuses on providing smart (intelligent)
system to students for learning. ITS is a system which is implemented on computers that uses Artificial Intelligence(AI)
techniques for advanced learning experience where human tutor is replaced by a virtual tutor. It presents education and other
study related information in a flexible and personalized way. An ITS mainly consists of three models Domain Model, Student
Model, Tutor Model.
Keywords: Affective Computing, Comparative Study, e-Learning, Intelligent Tutoring System (ITS), Models of Tutoring
System
_______________________________________________________________________________________________________

I. INTRODUCTION

An Intelligent Tutoring System (ITS) is an advanced computing system that produces customized and prompt instructions to
learners without requiring intervention of human teacher. Intelligent Tutoring Systems have a common purpose of providing a
learning experience in a meaningful and effective manner by using a number of computing technologies. The main focus of
Intelligent Tutoring System is on one-to-one learning and on training based on the individual preferences. With the help of ITS a
high quality of education can be provided.
Traditional ITS model consisted of i. The Domain Model ii. The Student Model iii. The Tutoring Model [1]. The Domain
Model includes the information/knowledge of the system i.e. how the tutor is being set up and how the student sees the
instructions. The domain consists of the properties of domain and the tasks to be trained. The Student Model is responsible for
the assessment of student’s knowledge and learning. It gives the information about student’s ability and expertise based on the
tests performed. The test result shows the improvement area and the area to work on. The Tutoring Model uses different
algorithms and approaches based on the analysis for the learning of the student.

II. LITERATURE SURVEY

Ramón Zatarain-Cabada, M. L. Barrón-Estrada et al. in [2] proposed ITS which used visual affect and learning style recognition
software systems. The First system implemented basic emotions in student’s expressions using Paul Ekman’s method and the
second recognizes the learning style based on the Felder-Silverman model. Both the systems are put together to form the ITS.
Minami Otsuki, Masaki Samejima in [3] proposed a case based ITS for e-Learning on project management. For teaching the
Project Manager (PM) typical situations encountered during a project for its successful completion. Cases are built from past
experiences such that PM gets to know the real time problems and how to deal with it.
David Arnau, Miguel Arevalillo-Herraez et al. in [4] proposed the paper for Arithmetic Problem Solving that works under the
supervision of a virtual tutor. This is based on the analysis of the thoughtful processes that takes place during problem solving
and tasks performed by a human while solving arithmetic problem under the supervision of one-to-one tutoring situation. ITS
analyses the student’s approach for solving problems and provides the solution based on previous response by user and the
problem type.
Patil Deepti Reddy and Dr. Sasikumar M. in [5] proposed paper that emphasized more on the student model, as the author
believed that knowledge state of the student help in generating lessons, problems, feedback and guidance in a personalised way.
This paper used ITS for the learning of natural language i.e. Telugu. Confidence level of a learner is calculated to find out how
much comfortable learner is in learning and answering questions based on this confidence level ITS works to improve the
weakness of the learner.
Prema Nedungadi and M.S.Remya in [6] proposed a ITS model PC-BKT (Personalized and Clustered Bayesian Knowledge
Tracing) which is the enhancement of BKT-PPC model. BKT-PPC model had few drawbacks as it was based on prior learning.
Prior Learning was common for every student irrespective of their capabilities. PC-BKT is the improved method with improved

All rights reserved by www.journal4research.org 7

 Comparative Analysis of Intelligent Tutoring System Approaches
(J4R/ Volume 03 / Issue 10 / 002)

accuracy. In the proposed model students with similar skills are clustered and then personalised learning is provided. The
Authors proposed an Algo for Personalised BKT Model.
Ika Widiastuti and Nurul Zainal Fanani in [7] proposed ITS which is based on adjustment of difficulty level of question in
order to provide an adaptive system. The adjustments are done automatically by using fuzzy logic and item response theory. The
architecture (Figure 1) includes Expert Model, Pedagogical Model, Student Model, and Communication Model and the
adjustment levels are set by varying Question Level and adjusting Learner Level.

Fig. 1: Structure of Adaptive Learning [7]

Tenzin Doleck, Ram B. Basnet et al. in [8] proposed a ITS model used for training novice physicians in diagnosing diseases
on virtual patient using Hidden Markov Models. Previous researches in Clinical Reasoning was focused on whether disease is
diagnosed correctly or not however in this paper author focused more on the actions and sequential steps that leads learner to
deduce the disease.
Danial Hooshyar, Rodina Binti Ahmad et al. in [9] proposed a flowchart based computer programming tutoring for the new
learners of programming and named the method as FITS (Flow-Chart Intelligent Tutoring System), for decision making
Bayesian network technique is applied for taking decisions and to respond according to the knowledge level of learner.
Therefore, the proposed system uses the dynamic analysis for learning.
Mohamed Yassine Si Allal Zarouk, Mohamed Khaldi in [10] authors exploited the use of self-regulated learning(SRL) based
on Zimmerman cyclic model to provide metacognitive based response by considering learner’s differences in skills, type of
knowledge learner want and metacognitive needs.
Nguyen Thai-Nghe and Lars Schmidt-Thieme in [11] emphasized on the Student Model to track information of individual
student in order to analyze student’s performance, such that system can provide early feedbacks to improve the learning results
of the student. The analysis done by ITS takes into consideration the time spent by learner for a particular problem, hints
requested, correct and incorrect answers. Authors proposed multi-relational factorization approach for student modelling.
Danial Hooshyar, Rodina Binti Ahmad et al. in [12] proposed a game based learning by using concepts of ITS to teach
introductory programming. A Game-based Intelligent Tutoring System (GITS) is developed in the form of competitive board
games. The board game uses the Snake and Ladder game to improve web based programming solving techniques, Bayesian
network is also used for the purpose of decision making.
Maria Sette, Lixin Tao et al. in [13] proposed that cyber learning have a drawback of personal and assessment driven learning
which leads to confused student. With huge and diverse learning materials and lack of focusing on specific concept rather than
general concept. The industry standard is to use the Wen Ontology Language(OWL) for representing the knowledge but It had
drawback as it used “first-class” relation, “is-a”, between the concepts and different knowledge areas need customized relations
among the concepts such as “part-of” and time dependency. Therefore, research provides an extension of OWL to Knowledge
Graph (KG) to represent knowledge model with custom relations.
Yu Yan, Kohei Hara, Takenobu Kazuma et al. in [14] proposed a technique SKP-LS (Syntax Knowledge Point based Learning
Status ) in order to overcome the demerits of existing computer systems like the inefficiency of computer to define learner’s
individual learning status as the as the information used is not detailed, the student model cannot be handled by automated
computer and to tackle all these problems authors proposed SKP-LS as SKP includes individual student’s learning status .
Sanae Zrigui, Abdelaziz Ait Moussa in [15] proposed a method to provide the student’s specific needs by automatically
providing compatible courses and giving analysis based on the past and current student’s profile. Authors proposed an adoptive
method to change the interface model such that tutor can update donnish knowledge and related pedagogical knowledge as per
the learner.

All rights reserved by www.journal4research.org 8

 Comparative Analysis of Intelligent Tutoring System Approaches
(J4R/ Volume 03 / Issue 10 / 002)

III. COMPARATIVE ANALYSIS OF INTELLIGENT TUTORING SYSTEM

Table - 1
ITS Techniques with Different Parameters
Techniques
Hidden
Parameters Flowchart
PC-BKT CBL Fuzzy Logic Markov
based ITS
Model
To help
Prediction of trainee
Adjustment
student’s Educating Project physicians To enhance
of student
performance on Managers based on learn programming
Application level and
the basis of former real time diagnose skills of novice
question
skill and learning scenarios/cases correctly on programmers
difficulty
ability of student virtual
patients
Programmer
Clustered student Students with
Novice having no prior
Target Audience having varying Project Manager different
Physicians knowledge of
learning ability levels of skills
programming
Case description of Student’s
Knowledge
previous project in the level(high, Learner’s Class lecture
Components in the
Data Set form of problems medium, low) action during notes and their
form of steps
faced in the previous based on Pre- diagnosis solutions
taken by student
project and answer set tests
Accuracy High High Medium High Medium
Performance High High Medium High Medium
Prior Knowledge required
Yes Yes No Yes No
for system
Capability Matrix
Algorithms/Methodologies K-Means Bayesian
Jaccard Index
to execute techniques Clustering Network
Algorithm
Pre-test is
Sparse Data
Limitation/Challenges always
Matrix
required

IV. CONCLUSION

Comparative Analysis of ITS is mainly based upon the various techniques used to implement it. Each technique has some
advancement, limitation, application, data set in comparison to other techniques. Techniques like Fuzzy Logic which is the most
basic technique for the implementation of intelligent systems is used to implement ITS where it is applied to adjust student’s
level by taking pre-tests and accordingly set the difficulty level of question. Similarly, Bayesian Network is used in several ITS
for improved predictions. Several other techniques have also been developed for the implementation of ITS and on the basis of
different parameters discussed in Table 1 many applications have been developed.

REFERENCES
[1] H. L. Burns, J. W. Parlett and J. B. Bushman, "Intelligent Tutoring Systems- Perspective," 1989.
[2] R. Zatarain-Cabada, M. L. Barrón-Estrada, J. O. Camacho and C. A. Reyes-García, "Integrating Learning Styles and Affect with an Intelligent Tutoring
System," Proceedings - 2013 12th Mexican International Conference on Artificial Intelligence, MICAI 2013, pp. 247-253, 2013.
[3] M. Otsuki and M. Samejima, "An intelligent tutoring system for case-based e-learning on project management," Proceedings - 2013 IEEE International
Conference on Systems, Man, and Cybernetics, SMC 2013, pp. 3471-3476, 2013.
[4] D. Arnau, M. Arevalillo-Herraez and J. A. Gonzalez-Calero, "Emulating Human Supervision in an Intelligent Tutoring System for Arithmetical Problem
Solving," IEEE Transactions on Learning Technologies, vol. 7, no. 2, pp. 155-164, 2014.
[5] P. D. Reddy and M. Sasikumar, "Student Model for an Intelligent Language Tutoring System," Proceedings - IEEE 14th International Conference on
Advanced Learning Technologies, ICALT 2014, pp. 441-443, 2014.
[6] P. Nedungadi and M.S.Remya, "Predicting Students’ Performance on Intelligent Tutoring System - Personalized Clustered BKT," Proceedings - Frontiers
in Education Conference, FIE, Vols. 2015-Febru, no. February, 2015.
[7] I. Widiastuti and N. Z. Fanani, "Adjustment Levels for Intelligent Tutoring System using Modified Items Response Theory," 2014 1st International
Conference on Information Technology, Computer, and Electrical Engineering: Green Technology and Its Applications for a Better Future, ICITACEE
2014 - Proceedings, pp. 217-221, 2014.

All rights reserved by www.journal4research.org 9

 Comparative Analysis of Intelligent Tutoring System Approaches
(J4R/ Volume 03 / Issue 10 / 002)

[8] T. Doleck, R. B. Basnet, E. Poitras and S. Lajoie, "Towards Examining Learner Behaviors in a Medical Intelligent Tutoring System: A Hidden Markov
Model Approach," Souvenir of the 2015 IEEE International Advance Computing Conference, IACC 2015, pp. 329-332, 2015.
[9] D. Hooshyar, R. B. Ahmad, M. Fathi, M. Yousefi and M. Hooshyar, "Flowchart-based Bayesian Intelligent Tutoring System for Computer Programming,"
2015 International Conference on Smart Sensors and Application, ICSSA 2015, pp. 150-154, 2015.
[10] M. Y. S. A. ZAROUK and M. KHALDI, "Towards an Intelligent Tutoring System developing Metacognition: Cyclical model," 2015 10th International
Conference on Intelligent Systems: Theories and Applications, SITA 2015, 2015.
[11] N. Thai-Nghe and L. Schmidt-Thieme, "Multi-Relational Factorization Models for Student Modeling in Intelligent Tutoring Systems," Proceedings - 2015
IEEE International Conference on Knowledge and Systems Engineering, KSE 2015, pp. 61-66, 2016.
[12] D. Hooshyar, R. B. Ahmad, M. Yousefi, M. Fathi, S.-J. Horng, M. Hooshyar and A. R. Dooraki, "Improving Web-based Problem Solving Skills Of
Students With A Novel Game-Based Intelligent Tutoring System," pp. 0-3, 2015.
[13] M. Sette, L. Tao, K. Gai and N. Jiang, "A Semantic Approach to Intelligent and Personal Tutoring System," Proceedings - 3rd IEEE International
Conference on Cyber Security and Cloud Computing, CSCloud 2016 and 2nd IEEE International Conference of Scalable and Smart Cloud, SSC 2016, pp.
261-266, 2016.
[14] Y. Yan, K. Hara, T. Kazuma and A. He, "A method for personalized c programming learning contents recommendation to enhance traditional instruction,"
Proceedings - International Conference on Advanced Information Networking and Applications, AINA, pp. 320-327, 2017.
[15] S. Zrigui and A. A. Moussa, "Automatic Insertion of Subject Content in Domain Knowledge of an Intelligent Tutoring System," 2017 Intelligent Systems
and Computer Vision (ISCV), vol. 00, no. , pp. 1-5, 2017, doi:10.1109/ISACV.2017.8054971, pp. 3-7, 2017.

All rights reserved by www.journal4research.org 10