6. Conclusions
This study proposes a conceptual map method for modeling the prerequisite relationships
among concepts to be learned. Based on the proposed model, an intelligent tutoring system is
implemented which can identify poorly-learned and well-learned concepts for individual students
and provide appropriate individual learning guidance to enhance learning performance. To evaluate the efficacy of the novel approach, an on-line experiment involving 60elementary school
students was performed. The experimental results reveal that the experimental group, i.e. the
group of students who received the learning guidance generated by the tutoring system, made
significant progress compared with the control group. Consequently, we conclude that the novel
approach helps students to improve their learning progress.
Although the experiment achieved positive results for a natural science course, it would be
interesting to know whether the same approach would work, and how well, for various other
kinds of courses, such as language courses, mathematics courses, science courses, engineering
courses, and social science courses. Consequently, further investigations have been planned to
apply the novel approach to on-line tutoring for different courses.
Another relevant topic is the development of a computer system to assist teachers in constructing item banks and concept effect relationships. One branch of our research group is
attempting to develop a concept relationship construction system, which utilizes statistical theory
and data mining technology to help teachers to identify relationships among concepts.
Developing an intelligent tutoring system clearly needs to consider various factors, including
maintaining the item bank, constructing the concept relationships, generating feasible test sheets,
implementing adaptive tutoring strategies and designing subject materials. Consequently, an
extended project of ITES, the PTTD (Personalized Tutoring, Testing and Diagnosis Environment) project, was initiated in August 2001, and is scheduled to be finished in July 2004. The
entire project is supported by National Science Council of Taiwan and aims to develop an integrated system that offers tutoring, testing and diagnosis functions. The conceptual map model
proposed in this study will be employed to identify student learning status and to provide learning
suggestions for individual students in the PTTD project.