<![CDATA[This study examines the effectiveness of deep learning as an adaptive learning method for elementary students with diverse learning styles at SD-IT Ass-Shiddiqin. It is based on the challenge that students exhibit varied learning preferences—visual, auditory, reading/writing, kinesthetic, and multimodal—requiring more personalized instructional approaches. Deep learning, as part of artificial intelligence, enables real-time analysis of student interaction data to identify learning patterns that are difficult to detect manually. This capability allows the system to deliver learning materials tailored to individual student needs. The research employed a quasi-experimental design using a non-equivalent control group approach. Two classes were involved: an experimental group that used deep learning-based adaptive learning and a control group that applied conventional methods. Instruments included a VARK questionnaire to identify learning styles, pre-tests and post-tests to measure learning outcomes, as well as observation and documentation. Data analysis involved tests of normality, homogeneity, t-test, and ANOVA.The findings revealed a significant improvement in the experimental group’s academic performance. Post-test scores increased markedly compared to pre-test results and were higher than those of the control group. The t-test indicated a significant difference (p < 0.05), while the effect size (Cohen’s d = 1.32) showed a very strong impact. ANOVA results also confirmed that learning styles significantly influenced adaptive learning outcomes (p < 0.05). Overall, deep learning-based adaptive learning proved effective in enhancing motivation, engagement]]>
