Cross-disciplinary Teaching for Chinese College Students Based on Enhanced Graph Neural Network - Taking Piano Class as an Example

Abstract

The aim of this study is to comprehensively understand and optimize the curriculum of interdisciplinary course settings in piano courses for Chinese university music students, with the goal of enhancing educational quality and student learning outcomes. The study proposes an intelligent piano teaching recommendation feedback system based on the Knowledge Graph Enhanced Graph Neural Network (KGEGNN). This system leverages deep learning techniques combined with knowledge graphs to extract features from user learning behaviors, thereby providing personalized teaching recommendations and real-time feedback. Additionally, it utilizes enhanced graph neural networks to extract reinforcement items from indirect interactions among users, aggregates embeddings of users and items, and makes recommendations through prediction functions to offer personalized music teaching suggestions and feedback. In terms of experimental results, the KGEGNN algorithm achieves 94.58% for Top1-Accuracy and 91.29% for Top1-F1 score, while achieving 95.36% for Top5-Accuracy and 92.81% for Top5-F1 score, representing at least a 6% improvement in prediction accuracy. Thus, the Intelligent Piano Teaching Recommendation System constructed in this study demonstrates significant advantages in multi-candidate prediction and recommendation accuracy, with important potential and practical value in enhancing learning outcomes and promoting curriculum optimization.