Application of Fuzzy Theory for Adaptive Learning Diagnosis System

This paper presents an application of Fuzzy Theory for Adaptive Learning Diagnosis System (FADS), which consists of five parts: calculating the concept similarity of different chapters; producing the personal adaptive reading mark automatically; calculating the learning degree of each chapter; constructing the fuzzy membership function; and evaluating the adaptive items selection mechanism. The system uses the similar concept method to compute the association between test items and teaching materials, and depends on a learner’s practice and test results when selecting and marking the important paragraphs from readings by bolding the text automatically. Each learner has personal adaptive marks on their teaching materials, used to identify the weak parts. Additionally, the system uses the associations and results from practices to compute the personal learning degree of each learner. FADS focus on the functions of a review and detailed explanation of answers, providing learners with diverse feedback and creating a personal self-study environment for each learner. The system transfers the difficulty of each chapter by using the fuzzy membership function to select appropriate items for the learner. A total of 200 fourth-grade students in six social studies classes participated. The learners were divided into three groups, two classes per group; the groups were the experiment group, the comparison group, and the control group. The experimental method was quasiexperimental. The experiment group used FADS for learning diagnosis and assisted learning; the comparison group used traditional error rates for learning diagnosis and assisted learning; and the control group did not use a system for learning diagnosis and assisted learning. The results indicate that the use of FADS to diagnose and assist learning enabled the students in the experimental group to perform better than the two other groups. Since FADS makes online learning diagnosis analysis more effective and user-friendly, it can be an adaptive and flexible learning diagnosis system.