International Journal of Artificial Intelligence & Applications (IJAIA), Vol.2, No.

2, April 2011

An Intelligent Tutoring System for Learning Java

Objects
S. Abu-Naser1, A. Ahmed1, N. Al-Masri1, A. Deeb2, E. Moshtaha2 and M. Abu-
Lamdy2
1
Faculty of Engineering and Information technology, Al-Azhar University, Gaza,
Palestine.
2
Faculty of Information Technology, Palestine University, Gaza, Palestine.
Corresponding author: Samy S. Abu-Naser, Faculty of Engineering and Information technology, Al-
Azhar University, Gaza, Palestine. Tel:(+9708)2824020 Fax:(+9708)2832907

ABSTRACT
The paper describes the design of a web based intelligent tutoring system for teaching Java objects to
students to overcome the difficulties they face. The basic idea of this system is a systematic introduction
into the concept of Java objects. The system presents the topic of Java objects and administers
automatically generated problems for the students to solve. The system is dynamically adapted at run
time to the student’s individual progress. The system provides explicit support for adaptive presentation
constructs. An initial evaluation study was done to investigate the effect of using the intelligent tutoring
system on the performance of students enrolled in computer science III in the Faculty of Engineering and
Information technology at Al-Azhar University, Gaza. The results showed a positive impact on the
evaluators.

KEYWORDS:
Intelligent Tutoring System, Java Objects, Problem Generation, JO-Tutor

1. INTRODUCTION
Intelligent Tutoring Systems (ITSs) can be traced back to the early 1970s, when Carbonell
tried to combine methods of Artificial Intelligence (AI) with Computer Aided Instruction
(CAI)[9]. Thus, the first generation of ITSs are more or less a kind of “intelligent” CAI. Their
main task is stated by Lelouche: “The basic principle of ‘intelligent’ CAI is that it should know
the taught material”[15]. Knowledge about the taught material is embedded in the ITS in form
of expert systems, that is, the expert module [3,11]. The integration of insights of cognitive
science in ITSs, has led to what today is called an Intelligent Tutoring System[2]. In addition to
the knowledge about the taught material, these systems have knowledge about pedagogical
strategies and knowledge about the student, realized as pedagogical module and student
module, respectively.
The classical ITS architecture, first described by Clancey, consists of the components: expert
module, pedagogical module, student module, and user interface[10]. The naming of the
components varies. Sometimes, depending on the training domain, a component for automatic
generation of exercises is also part of the ITS. Whereas the ITS’s constituents seem to be part
of common agreement in the ITS community, the role and the functionality of each of the
components varies a lot[1,3]. The reason for this can partly be seen in the different application
domains. One can easily imagine that training in mathematics places different demands on ITSs
than clinical medicine training. Another reason might be the realization of different learning
DOI : 10.5121/ijaia.2011.2205 68
 International Journal of Artificial Intelligence & Applications (IJAIA), Vol.2, No.2, April 2011

theories in ITS. Case-based learning, described in the former section, places special demands
on an ITS that are somewhat different in problem-oriented learning. These reasonable aspects
often necessarily lead to heterogeneous and incomparable systems. But even in the same
application domain and based on the same learning style, ITSs are often not comparable and
based on a complete different interpretation of the same architecture. Moreover, regarding only
the ITS architecture on a more abstract level, it becomes hard to find reasons for heterogeneous
realizations at all. A mixture of content and delivery functions, which is seemingly not based on
insights of research but on traditions of ITS development, can be found in ITS realizations.

In other ITSs, the expert knowledge base is a simple database without own functionality[17].
The same situation can be found in the different ways the pedagogical knowledge module is
realized and embedded in the ITS. Thus, there are ITSs that consist of a set of interacting and
more or less separate subsystems, and there are ITSs consisting of passive components plus a
component that encapsulates the execution. Execution in this context contains the interaction
with the student, the evaluation of the student’s behavior and success, and the provision of
contents and navigation. Thus, two perspectives on the same architecture can be found,
reflecting different interpretation of the same modules:
ITS architecture consists of either separate independent subsystems or passive components with
centralized execution system.
Both perspectives have their advantages and disadvantages. However, the main system’s
philosophy regarding the realization of the components should be made clear to provide for
comparability of ITSs and reusability of ITS components. The advantage of this approach is
that the central steering component might be reused in different ITSs, as it is obviously
separated from the databases and the user interface. Figure 1 shows the suggested ITS
architecture with the tutoring process module is sketched.

Figure 1. ITS architecture with tutoring process module

69
 International Journal of Artificial Intelligence & Applications (IJAIA), Vol.2, No.2, April 2011

The design of the Intelligent Tutoring System for Learning Java Objects (JO-Tutor) adopted the
ITS architecture with automatically generated problems module.

The objectives of the JO-Tutor are:

• To build an intelligent tutoring system for problems for which the answers might not be
always quantitative

• To be able to generate unlimited number of problems automatically, thereby providing

as much practice with problem-solving as the student needs.

• To build familiar interactive interfaces like OS desktops

• To have a system that is dynamically adapt at run time to the student’s individual
progress

2. JO-Tutor Design
Over the years a few innovative tutoring tools have been studied in an attempt to improve the
quality, flexibility and cost effectiveness of teaching and learning. Kashy developed a tutor
called CAPA to help students in solving Physics problems[12].
Barker have developed a tutor for homework assignment in electronic and control system
discipline[6].
Bridgeman developed Interactive tutors like: PILOT and SAIL[7,8]. PILOT was for Learning
and grading. PILOT is a problem generation tool for graph algorithms, while SAIL is a LaTeX-
based scripting tool for problem generation.
JO-Tutor is unique with respect to the previous work in the following manner:
• The system was built to look like the most familiar interactive interfaces like OS desktops,
including (icons, drag and drop features, drop down menus, and pop up windows) which
are all integrated in one single window.

• The intelligent tutoring system was built for problems for which the answers might not be
always quantitative [1,2].

• WebToTeach is related to JO-Tutor[5], but it administers instructor pre-prepared problems,

and does not generate the problems automatically. JO-Tutor can generate unlimited number
of problems automatically, thereby providing as much practice with problem-solving as the
student needs.

Kashy have shown that the use of problem generation systems have increased student
performance by 10% in Physics, largely due to limitless time spent on the task[13].

JO-Tutor is designed to help students learn Java Object concepts by:

1. Gradually teaching the Java Object material to the students. The system is supported
with a student controlled voice narrator, which acts as a facility during learning.
2. Repeatedly solving automatically generated problems and
3. Obtaining the proper feedback.

70
 International Journal of Artificial Intelligence & Applications (IJAIA), Vol.2, No.2, April 2011

The JO-Tutor is designed to be used as a supplement to the traditional method of teaching (Java
textbook and Instructor), either during a laboratory session, after class training, or for
homework assignments.

JO-Tutor has the following modules: Pedagogical Module, Problems Generation Module,
Expert Module, Student Module, and Tutoring process module.

2.1 JO-Tutor Pedagogical Module Design

It has been noticed that students are having difficulties in understanding the concepts of Objects
in Java. To overcome these difficulties, an Intelligent Tutoring System for teaching Java
Objects called JO-Tutor have been developed to students enrolled in Computer Science III in
the faculty of Engineering and Information Technology at Al-Azhar University in Gaza. JO-
Tutor gradually introduces students to the concept of Java objects and automatically generates
problems for the students to solve. The key sections that draws the main structure of the
tutoring material[18,19,20] are:

3. Classes and Objects

• Classes, Objects, Methods and Instance Variables
• Declaring a Class with a Method
• Instantiating an Object of a Class
• Declaring a Method with a Parameter
• Instance Variables, set Methods and get Methods
• Primitive Types and Reference Types
• Initializing Objects with Constructors
• Access modifiers
• Composition
• Enumerations
• Static Class Members
• Final Instance Variables

4. Inheritance
• Superclasses and Subclasses
• protected Members
• Relationship between Superclasses and Subclasses
• Constructors in Subclasses
• Nested Classes

5. Polymorphism
• Introduction to Polymorphism and Examples
• Demonstrating Polymorphic Behavior
• Abstract Classes and Methods
• Final Methods and Classes

2.3 Problems Generation Module of JO-Tutor

Limited number of problems has been recently acknowledged as a potential shortcoming of
encoding a finite set of problems into a tutor [16]. A method used in literature to dynamically
generate problems is by using BNF-like grammar [14]. In this method, problems are generated
71
 International Journal of Artificial Intelligence & Applications (IJAIA), Vol.2, No.2, April 2011

by randomly instantiating templates written in the grammar. Each template can be carefully
designed with specific pedagogical objectives in mind.
Problem Generation Module is responsible for generating the template of code, theses
templates cover the main topics of Java Object Oriented Programming which are (functions,
classes, Inheritance & polymorphism). The module depends on randomly structuring of the
pieces of codes which forms the templates that consist of (access modifiers types, return types,
arguments data types, classes, methods, and arguments names). These structures are imported
from pre defined lists of keywords.
The template is generated with a previously intended problem each time it is requested,
followed by a related question and possible solutions. The questions have different styles
including either asking the student to correct a Java code, write a Java code, multiple choice, or
true/false. From the perspective of enhancing high availability of required tools for student
during learning process, the system provide a simple editor linked with a Java compiler, to
enable the user to test some pieces of Java code he wants, so that there is no need to get out of
the system environment to compile Java codes.
2.4 Expert Module of JO-Tutor
Expert Module was implemented to gather the necessary information for generating the
feedback[4]. The expert module is capable of solving the generated problems by parsing the
template. Since the expert module can execute any code, it can generate the correct answer for a
problem on its own, and determine whether the user’s answer is correct/incorrect. In addition to
whether the user’s answer is correct/incorrect, the module can provide the student with the
correct answer when it is requested. Furthermore, the module provides the student the proper
feedback in response to the student's answer.
2.5 Student Module of JO-Tutor
A new student must create his own account to have a profile. The profile has information about
the student such as his name, dates of login, score of each session, and learning progress during
the each session. The student's score can be viewed at any time during the session as a 3D Bar
chart that describes the student performance in solving problems in the following subjects:
casting, classes and inheritance.
2.6 Tutoring process module of JO-Tutor
Tutoring process module works as a coordinator that controls the functionality of the whole
system.

3.JO-Tutor User Interface Design

Figure 2 shows the user interface of the tutor that consists of icons and drop down menus. Once
the main screen is shown up, another screen is shown to enable the user to login into the
system. If the user is using the system for the first time, he should create a new username and
password (See Figure 3). After the user enters the correct username and password the system
gives control to the user of the main screen. From the main screen, the user can click on the
clock icon for generating questions(true/false, multiple choice, and correct Java code questions
(see figure 4)), the chicken icon for learning the material of Java Object where the user choose
the topic to learn(see figure 5), wheel icon for help about the system, notes icon for personal
notes, question mark icon shows the work team, editor icon for typing and compiling java
codes(see figure 6), and Pie chart icon for showing the user current scores in casting, classes
72
 International Journal of Artificial Intelligence & Applications (IJAIA), Vol.2, No.2, April 2011

and inheritance(see figure 7). Furthermore, the user can use the pop up menu to select any of
the functions of the icons.

Figure 2: The main user interface

Figure 3: The login screen

73
International Journal of Artificial Intelligence & Applications (IJAIA), Vol.2, No.2, April 2011

Figure 4: Different types of questions

Figure 5: Learning material of Java Object

74
 International Journal of Artificial Intelligence & Applications (IJAIA), Vol.2, No.2, April 2011

Figure 6: Editor and compiler menu

Figure 7: Final Results

4.Evaluation of the JO-Tutor

An initial evaluation of the JO-tutor was carried out by the lecturers and their students who
enrolled in Computer Science III (advanced Java) during the Fall semester of 2010/2011 in
Faculty of Engineering and Information Technology at Al Azhar University-Gaza. A questioner
consisting of the items in table 1 was filled out by each evaluator (Lecturers and Students). A
group of 3 lecturers and 20 students participated in the evaluation of the system. Table 1 shows
the overall rating of the lecturers and students who evaluated the system.

75
 International Journal of Artificial Intelligence & Applications (IJAIA), Vol.2, No.2, April 2011

Table 1: Shows the rating of the of the JO-Tutor by lecturers and students
Rating %
Item
Item
# Lecturers Students
1 The Quality of the JO-Tour Design 88% 92%
2 The importance of the topic covered (advanced Java) 94% 98%
3 Would you benefit from using the JO-Tutor? 86% 98%
4 Do you recommend using JO-Tutor for Computer Science III 100% 100%
(advanced Java) as a supportive tool?
5 Would you like to see similar tutoring system in other 100% 100%
courses?
Form the summary of Table 1, the evaluation of the JO-Tutor showed a positive impact on the
evaluators (Lecturers and students). Furthermore, they recommend that similar systems for
other courses to be implemented.

5. CONCLUSIONS AND FUTURE WORKS

The design of an Intelligent Tutoring System called JO-Tutor was described in this paper. JO-
Tutor was designed for teaching Java objects to students to overcome their difficulties. JO-
Tutor presents the topic of Java objects to the student and administers automatically generated
problems for him to solve. JO-Tutor is dynamically adapted at run time to the student’s
individual progress. An initial evaluation of JO-Tutor was carried out by the lecturers and
students taken the advanced Java course in the faculty of Engineering and Information
Technology at Al Azhar University in Gaza. The outcome of the evaluation was positive and
suggested that other intelligent tutoring systems be designed for other courses. We recommend
a comprehensive evaluation of the system to be carried out next time the course is offered.

REFERENCES
[1] Abu-Naser, S., 2009. Evaluating the effectiveness of the CPP-tutor, an
intelligent tutoring system for students learning to program in C++. J. Applied
Sci. Res., 5: 109 -114
[2] Abu-Naser, S., 2008. Developing an intelligent tutoring system for students
learning to program in C++. Inf. Technol. J., 7: 1055 –1060
[3] Amalathas, S., A. Mitrovic, R. Saravanan and D. Evison, 2010. Developing an
intelligent tutoring system for palm oil in ASPIRE. Proceedings of the 18th
International Conference on Computers in Education, Nov. 29-Dec. 3, Asia-
Pacific Society for Computers in Education, Putrajaya, Malaysia, pp: 101-103
[4] Anderson, J.R., A.T. Corbett, K.R. Koedinger and R. Pelletier, 1995. Cognitive
tutors: Lessons learned. J. Learn. Sci., 4: 167-207.
[5] Arnow, D. and O. Barshay, 1999. WebToTeach: An interactive focused
programming exercise system. Proc. Annu. Frontiers Educ. Conf., 1: 12A9/39 -
12A9/44
[6] Barker, D.S., 1997. CHARLIE: A computer-managed homework, assignment
and response, learning and instruction environment. Proc. Annu. Frontiers Educ.
Conf. Teach. Learn. Era Change, 3: 1503 - 1509
76
International Journal of Artificial Intelligence & Applications (IJAIA), Vol.2, No.2, April 2011

[7] Bridgeman, S., M.T. Goodrich, S.G. Kobourov and R. Tamassia, 2000. PILOT:
An interactive tool for learning and grading. Proc. SIGCSE Tech. Symp.
Comput. Sci. Educ., 32: 139-143.
[8] Bridgeman, S., M.T. Goodrich, S.G. Kobourov and R. Tamassia, 2001. SAIL: A
system for generating, archiving and retrieving specialized assignments using
LaTeX. Proceedings of the 31st SIGCSE Technical Symposium, Mar. 7-12,
Austin, TX, pp: 300-304
[9] Carbonell, J., 1970. AI in CAI: An artificial-intelligence approach to computer-
assisted instruction. IEEE Trans. Man Mach. Syst., 11: 190 - 202
[10] Clancey, W., 1984. Methodology for Building an Intelligent Tutoring System.
In: Methods and Tactics in Cognitive Sciences, Kintsch, W., J.R. Miller and
P.G. Polson (Eds.). Lawrence Erlbaum Associates, Hillsdale, New Jersey,
London, pp: 51-84
[11] Fournier-Viger, P., R. Nkambou and E. Mephu, 2010. Building Intelligent
Tutoring Systems for Ill-Defined Domains. In: Advances in Intelligent Tutoring
Systems, Nkambou, R., R. Mizoguchi, and J. Bourdeau (Eds.). Springer-
Verlag, Berlin, Heidelberg, pp: 81-101.
[12] Kashy, E., B.M. Sherrill, Y. Tsai, D. Thaler, D. Weinshank, M. Engelmann
and D.J. Morrissey, 1993. CAPA-An integrated computer-assisted personalized
assignment system. Am. J. Phys., 61: 1124-1130.
[13] Kashy, E., M. Thoennessen, Y. Tsai, N.E. Davis and S.L. Wolfe, 1997. Using
networked tools to enhance student success rates in large classes. Proc. Annu.
Frontiers Educ. Conf. Teach. Learn. Era Change, 1: 233-237
[14] Koffman, E.B. and J.M. Perry, 1976. A module for generative CAI and
concept selection. Int. J. Man Mach. Stud., 8: 397-410.
[15] Lelouche, R., 1999. Intelligent tutoring systems from birth to now. Kunstliche
Intelligenz, 13: 5-11.
[16] Martin, B. and A. Mitrovic, 2000. Tailoring feedback by correcting student
answers. Proc. Intell. Tutor. Syst., 1839: 383-392.
[17] Tang, Y., 2005. Qualitative reasoning and articulate software. Inform.
Technol. J., 4: 184-188.
[18] Daniel Liang, Y., 2010, Introduction to Java Programming, Comprehensive
(8th Edition), Prentice Hall.
[19] Deitel, P. and H. Deitel, 2009, Java How to Program: Early Objects Version
(8th Edition), Prentice Hall.
[20] Reges, S. and M. Stepp, 2010, Building Java Programs: A Back to Basics
Approach (2nd Edition), Addison Wesley.

77