Enhancing opera vocal education through advanced machine learning algorithms: analytics for talent development and curriculum design
Keyong Wang, Pinqi Zhang · Interactive Learning Environments · 2025
The goal of this study was to examine the effectiveness of a new system for providing personalized recommendations on educational materials in the learning process. The system was designed using machine learning models, specifically a long short-term memory deep neural network for recognizing the emotional tone of musical fragments in audio form and a recurrent neural network with a hierarchical attention mechanism for classifying vocal fragments by performance complexity. This system was employed to increase the flexibility of the learning process and enhance instructors’ ability to select relevant materials. The study involved a total of 118 opera students divided into experimental and control groups. The experimental group utilized the proposed system to select materials for one-to-one vocal lessons with an instructor. The results were compared using non-parametric Mann–Whitney tests for independent samples and Wilcoxon signed-rank tests for paired samples, revealing significant improvements in motivation components. In-depth interviews with 23 students provided detailed insights into their experiences with the system. Although most students expressed interest in the new technology, they identified several challenges. These challenges included a lack of accuracy and relevance in the results obtained and the complexity of the emotion classification system used.