![](bg6.png)
2.5.2 Instructor Interaction with EG
The second part of the “Interface” module is designed
as a tool for the instructors. They can initiate the
student model of a student and view their students’
progress. Initiating the student model of a student cor-
responds to recording his/her personal information.
3 COMPARISON OF EG WITH
OTHER EQUATION SOLVING
TUTORS
This section compares Equation Guru with other
equation solving tutors. For this comparison only
equation solving tutors in the domain of linear equa-
tions are considered.
Some equation solving tutors’ domains are re-
stricted to equations in some specific form. For in-
stance, Equation Solving Tutor (EST) (Ritter et al.,
1995) helps students in solving linear equations only
in the form of ax+b=c, while in E-Sit (Prince,2004),
the equations are only in the form of ax+bx=c. An-
other tutor, E-Tutor (Razzaq et al., 2004) supports
cross multiplication and expansion in the domain
of linear equations while AlgeBrain (Alpert et al.,
1999) has these features in the domain of linear and
quadratic equations. Different from these tutors, EG
supports all forms of linear equations, including equa-
tions with fractions. Furthermore, the format of equa-
tions is not restrictive - anything is accepted, as long
as it represents a valid equation.
Motivation and attracting the attention of the
learner are important aspects of the learning process.
In order to motivate the student, E-Sit uses a game.
The game window appears automatically whenever
the student gets a specific mark from the posed ques-
tions and the duration of the game depends on the stu-
dent’s success. AlgeBrain uses a character in its tutor-
ing process. Animation features of this character are
limited and it has no speech capability. EG uses Mi-
crosoft Agents for motivating and attracting the atten-
tion of the student which provides a highly effective
learning environment. These characters have a wide
range of animations and they can speak and easily en-
gage the student to the learning process.
E-tutor tutors in a dialog-based manner. Similarly,
the “Tutorial” part of EG is designed in this manner.
In all of the above mentioned equation solving tu-
tors, next problem selection is based on the informa-
tion available in the student model. EG also works
this way. In E-Sit, however, next problem selection
is based on a utility function which does not support
exactly student dependent tutoring.
AlgeBrain is a web-based and collaborative equa-
tion solving tutor while other tutors, including EG are
standalone applications.
Hint messages in Cognitive Tutor (Koedinger et al.,
2000) are generated sequentially to the student. It al-
ways provides a strong hint, by telling exactly what
to do at the end of the sequence. But this approach
contradicts with the principles of effective teaching
identified in (VanLehn et al., 1998). That is, the tu-
tor should not provide strong hints for the solution
of equations when students need them. If so, then
they may miss the opportunity to learn how to solve
equations when they are provided an answer and not
allowed to reason for themselves. EG generates an
appropriate hint message to the student when needed
but this message will never be strong.
In E-Sit, the next expected action (next step) from
the student is specific. For example, for the equation
3x-5=15, the next expected action is 3x=15+5. Also,
the name of the solution step (like transformation, ad-
dition, etc) must be submitted to the system by the stu-
dent. The correctness analysis of the student’s action
is based on this assumption. In such a system, if the
student submits a correct solution step ahead of the
expected one (x=20/3, for the above example), then
the ITS will consider this unexpected step as wrong
solution step. Therefore, the wrong evaluation will
yield a wrong student model. Furthermore, wrong
modeling will yield wrong pedagogical decisions and
strategies. In AlgeBrain, the next expected action is
a set of possible actions that can be applied to sim-
plify the considered equation. This discussion brings
us to a major point of strength in EG, when compared
to other equation solving tutors. In EG, the expected
solution step from the student is not specific. As long
as the student’s action results in an equation with the
same solution as the equation provided by the system,
it is assumed to be a correct move, and the student is
allowed to carry on. However, the student is warned if
his/her solution step takes him away from obtaining a
correct answer. Also, there is no need for the student
to provide the step name to the system.
Furthermore EG supports linear equations in a va-
riety of forms, including those with fraction additions
(this feature is missing in many other tutoring sys-
tems), and a blackboard at the right top corner of the
screen where solutions are displayed, creating a fa-
miliar medium, similar to a classroom learning envi-
ronment.
4 CONCLUSION AND FUTURE
RESEARCH DIRECTIONS
In this paper, we described the implementation strate-
gies of an intelligent tutoring system, called Equation
Guru (EG), which is designed to help high school
students at grade 8 with algebra. EG is an ITS that
IMPLEMENTATION STRATEGIES FOR “EQUATION GURU” - A User Friendly Intelligent Algebra Tutor
63