Smart tutoring: Subject with targeted tutor matching and real-time emotion engagement analysis

ABSTRACT: This research investigates the development of a “smart tutoring” system that incorporates real-time emotion recognition and personalized tutor matching to enhance learning experiences. The system aims to address diverse student needs by offering flexible and adaptive learning environments. Key features include automated scheduling, virtual tutoring with emotion detection, and personalized tutor recommendations, utilizing Convolutional Neural Networks (CNN) for emotion recognition and a content-based filtering algorithm for tutor matching. The system monitors students emotional states and learning preferences, allowing for real-time adjustments in teaching strategies. Evaluated using ISO 25010 criteria, the system showed positive results, with mean scores ranging from 3.82 to 4.33 on a 5-point Likert scale. While the system met its core objectives, there remains potential for further improvement. This study demonstrates the potential of using CNN and content-based filtering algorithm to create more personalized, emotionally responsive, and adaptable learning environments.

Filipiniana