PROFILING K-12 LEARNERS IN SMART E-LEARNING: A COMPARATIVE STUDY OF K-MEANS AND FUZZY C-MEANS USING THE TAM, TTF, AND SUS FRAMEWORKS

The rapid rise of smart technologies in education has reshaped STEM and coding learning,promoting adaptive anddata-driven learning for diverse K-12 students. This study integrates the Technology Acceptance Model (TAM), TaskTechnology Fit (TTF), and System Usability Scale (SUS) to analyze students’ perceptions of usability, motivation,satisfaction, and learning outcomes. Data were collected from 450 students across elementary, middle, and highschool levels to explore variations in digital learning experiences. Clustering analysis using K-Means and Fuzzy CMeans was applied to identify learner profiles based on perception and interaction patterns. The results show thatFuzzy C-Means produced more interpretable and coherent clusters than K-Means, capturing smoother boundariesand overlapping learner characteristics. Three profiles emerged: Young Explorers, Motivated Builders, andIndependent Coders, reflecting developmental transitions across educational levels. These findings highlight howclustering approaches can enhance understanding of learner diversity and inform adaptive design in e-learning. Thestudy contributes to personalized learning analytics by linking behavioral perspectives with computational modeling,offering insights for developing inclusive and user-centered digital education systems.