Innovative models for enhanced student adaptability and performance in educational environments.

Area of Science:

• Educational Data Mining
• Machine Learning in Education
• Adaptive Learning Systems

Background:

• Traditional educational systems face challenges in adapting to diverse student needs.
• Predictive analytics can enhance student performance evaluation and feedback mechanisms.
• Existing models may lack the precision required for nuanced analysis in flexible learning environments.

Purpose of the Study:

• To forecast student learning adaptability.
• To predict and evaluate student academic performance.
• To utilize aspect-based sentiment analysis for detailed student feedback insights.

Main Methods:

• Data preparation and balancing techniques were employed.
• Feature extraction methods were utilized for pattern identification.
• The WideResNeXT architecture was optimized using the Modified Jaya Optimization Method, resulting in the WResNeXt-MJ model.

Main Results:

• The WResNeXt-MJ model demonstrated exceptional performance across datasets.
• Achieved an average accuracy of 98%.
• Reported a log loss of 0.05% and a precision score of 98.4%.

Conclusions:

• The WResNeXt-MJ model offers a robust solution for enhancing predictive capabilities in flexible learning environments.
• The model significantly improves student performance prediction and feedback analysis.
• This approach is suitable for global higher education contexts.