<![CDATA[Learning algebra is often challenging for high school students due to its abstract nature, leading to loss of motivation. This study designed and implemented an Adaptive AI Platform that personalized practice problems based on real-time performance analysis and student interests. A quasi-experimental study with 120 10th grade students demonstrated a significant effect of the platform on improving academic achievement. The experimental group (n=60) using the platform demonstrated significantly higher TPA posttest scores (adjusted mean: 81.95) than the control group (n=60) (adjusted mean: 73.65), after controlling for pretest scores. In addition, the platform substantially increased students’ learning motivation and self-confidence (Cohen’s d = 2.48 for the experimental group), supported by instant feedback and difficulty adaptation. The integration of Deep Knowledge Tracing (DKT) and Item Response Theory (IRT) models enabled dynamic content adjustments aligned with students’ cognitive progress. The results were also supported by students’ positive perceptions regarding the engagement and effectiveness of the platform. This study recommends the integration of adaptive AI platforms in high school mathematics education with adequate teacher support and infrastructure, and suggests future research for longer intervention durations and ethical analysis of AI.]]>
Copyrights © 2025
