A deep learning-based model for evaluating elderly education quality from the perspective of intelligent education

Elderly education quality assessment, involving diverse and multi-modal data, has significantly advanced with the adoption of deep learning techniques. This study presents an innovative hybrid CNN-LSTM model that combines the spatial feature extraction capabilities of Convolutional Neural Networks (CNNs) with the sequential pattern analysis strengths of Long Short-Term Memory (LSTM) networks. By processing structured data such as attendance and test scores alongside unstructured inputs like learner feedback and behavioral observations, the model delivers a comprehensive evaluation framework. Furthermore, integrating smart education principles, optical imaging technologies such as panoramic cameras and multispectral imaging are leveraged to capture key classroom environmental factors, including lighting, seating arrangements, and equipment usage, enriching the dataset and boosting evaluation accuracy. Tested on a real-world dataset from elderly universities, the model achieved a precision of 91.2%, outperforming traditional methods like logistic regression and random forest algorithms. It also demonstrated enhanced F1-scores and resilience to imbalanced data, effectively addressing the primary challenges in elderly education assessment. Attention-based feature analysis further revealed that learner satisfaction, engagement, and environmental conditions are critical to educational outcomes, enhancing the model’s interpretability. By aligning with modern smart education frameworks, this research offers a robust and practical solution for improving elderly education quality, showcasing the effectiveness of integrating advanced deep learning techniques with optical imaging to manage complex datasets.