The development of association rules for student performance analysis using FP-Growth algorithm as a guideline for multidisciplinary learning

Abstract

This study aims to develop association rules for student performance analysis using the FP-Growth algorithm. The data used for developing the association rules comprised 107 student reports. The reports, voluntarily provided by 107 junior high school students, consisted of student achievement results across 8 subject areas: Thai Language, Mathematics, Science, Social Studies, English Language, Computing Science, Visual Arts, and Home Economics. The data was applied to developing association rules using the FP-Growth algorithm towards WEKA, a machine learning software. The research team designed the process consisting of the following 5 stages: data collection, data preparation, model formulation, model evaluation, and model application. After achieving the association rules, the research team applied them to the prototype development of a student performance analysis system for promoting students' academic excellence. The system could be operated by Android mobile phones. According to the research results, the association rules developed by the algorithm provided a confidence level of 92%, and a rule of 7 rules will be generated. The findings indicated the correlations between the subject areas, which shared similar individual students' academic achievements (≥ 80 scores). The association rules could be applied to the multidisciplinary curriculum planning, which benefited students and promoted academic excellence. For example, by applying Rule, it could be assumed that students who earned 80 scores or higher in the English subject would likely earn identical scores from their Thai Language class. Therefore, they could effectively learn to integrate English and Thai languages. To illustrate, students may be asked to translate song lyrics from English to Thai, serve as tourist guides or translators, or even give welcome speeches to foreign guests.