A deep learning-based adaptive e-module in mathematics: Design and evaluation for visual communication design vocational students

Purpose: This study aims to design, implement, and evaluate a deep learning-based adaptive e-module to teach mathematical concepts. Specifically, the topic of optimum value to tenth-grade vocational students in the Visual Communication Design (DKV) program. The goal is to determine the module's feasibility, practicality, and effectiveness in improving students' problem-solving skills and learning motivation.Method: The research employed the ADDIE development model, consisting of five stages: analysis, design, development, implementation, and evaluation. Validation was carried out by three experts in education and technology. The product was tested on 36 vocational students, using instruments such as expert validation forms, Likert-scale questionnaires, pre- and post-tests, interviews, and classroom observations.Findings: Expert evaluations indicated high feasibility with scores ranging from 4.2 to 4.6. The average student score increased from 57.97 (pre-test) to 69.44 (post-test), while the standard deviation decreased from 11.315 to 8.914. A low correlation between pre- and post-test scores (r = –0.022; p = 0.899) suggests that the module was equally effective for students across different ability levels. Feedback from both teachers and students confirmed the module’s high practicality and engaging nature.Significance: The findings demonstrate that a deep learning-based adaptive e-module can enhance visual and contextualized mathematics learning for vocational students. It promotes conceptual understanding, learning motivation, and problem-solving skills regardless of initial performance. This study highlights the transformative potential of integrating artificial intelligence into instructional design, especially for creative-industry-focused education programs.