Development and Application of Convergence Education about a Support Vector Machines for Elementary Learners

Abstract

This paper proposes an artificial intelligence (AI) convergence education program that teaches the concept and principles of support vector machines (SVM). It was applied to elementary school students to check the effectiveness of education. The developed program describes the decision boundaries and margins of the SVM through the distance between the vertical and parallel lines of the elementary mathematics curriculum. The SVM education program was piloted to 18 3rd graders and 16 5th graders. Students' reactions were observed during classes, and learning output and self-evaluation results were analyzed after the classes. As a result, most students intuitively inferred the location of the trails, which represents the decision boundary. At this time, the overall activity performance accuracy of fifth-grade learners was higher, and the rate of reasonable inference on the principle of setting was also higher. The completion of the fourth-grade math curriculum also affected the students’ understanding. However, in the self-evaluation of the understanding of learning content, contrary to the actual comprehension, the average value was higher in the third grade. This was because middle-grade learners had a greater tendency to feel satisfaction and achievement when they newly learned about the unfamiliar AI principle. However, higher-grade learners presented more meaningful post-class. The novelty of this paper can be found in that it attempted to educate the SVM for elementary learners. In the future, it is necessary to apply the program to a larger number of learners, supplement and develop the program and contribute to the AI principles education for elementary school students. Keywords: support vector machines, artificial intelligence education, convergence education, elementary education, artificial intelligence algorithm. DOI: https://doi.org/10.55463/hkjss.issn.1021-3619.60.75