Vocal Course Quality Evaluation Model Using Hybrid Neural Networks and Meta-Learning Optimization

This paper proposes a vocal course quality evaluation model, DLM-Net, based on the integration of neural networks and meta-learning, to address the shortcomings of existing evaluation methods in handling data diversity and complexity. The DLM-Net model combines Deep Neural Networks (DNN), Long Short-Term Memory Networks (LSTM), and meta-learning optimization strategies. The DNN module extracts high-level features from audio data, the LSTM module captures temporal dependencies, and the meta-learning module enhances the model’s generalization ability. Experimental results on the GTZAN and VoxCeleb datasets show that this model significantly outperforms other comparison models in terms of Accuracy, Recall, F1-Score, and AUC. The DLM-Net model demonstrates strong adaptability and classification ability in diverse data scenarios, offering a new technological approach and research direction for vocal course quality evaluation, and making a positive contribution to the intelligent development of the education field.