SmartFlex Learning Ecosystem: Integrating AI and Learning Analytics in Vocational Mathematics Education

Abstract

Background The increasing adoption of artificial intelligence and learning analytics has accelerated the transformation of vocational education toward more adaptive and data-informed learning environments. However, existing approaches are often implemented as isolated technologies rather than integrated learning ecosystems. This study developed and evaluated a SmartFlex Learning Ecosystem that combines artificial intelligence, learning analytics, and flexible learning design to support vocational mathematics education. Methods A Design-Based Research approach was employed through four iterative phases: analysis, design, implementation, and evaluation. The SmartFlex Learning Ecosystem was implemented over a 12-week period involving 172 vocational students. Quantitative data were collected through structured questionnaires and learning analytics records, while system effectiveness was examined using Partial Least Squares Structural Equation Modeling and implementation-based learning analytics. Results The findings indicate that the SmartFlex Learning Ecosystem positively influenced learning engagement (β = 0.62, p < 0.001) and adaptive learning (β = 0.58, p < 0.001). Artificial intelligence-driven personalization showed a strong association with adaptive learning (β = 0.65), while learning analytics capability contributed to learning engagement (β = 0.53). Learning engagement (β = 0.61) and adaptive learning (β = 0.64) were positively related to mathematical competence, which subsequently influenced vocational skill readiness (β = 0.67). The model demonstrated substantial explanatory power (R 2  = 0.71) and predictive relevance (Q 2  = 0.52). Implementation data further showed an increase in login frequency of 142% and an improvement in assignment completion rates from 64% to 88%. Conclusions The SmartFlex Learning Ecosystem demonstrates the potential of integrating artificial intelligence, learning analytics, and flexible learning strategies within vocational mathematics education. The findings suggest that data-informed and adaptive learning environments can support student engagement, mathematical competence, and vocational skill readiness while providing a framework for future development of technology-enhanced vocational learning systems.