Efficient music analysis mechanism based on AI and IoT data mining

Minglong Wang,Daohua Pan
DOI: https://doi.org/10.1002/itl2.436
2023-05-16
Internet Technology Letters
Abstract:Chinese culture is depicted in a profound manner through opera music. With the advancements in deep learning and IoT technology, numerous studies have increasingly utilized neural networks to supersede conventional acoustic models. This paper explores the emotion classification of Qinqiang Opera through the utilization of cutting‐edge research methods. Firstly, we improve the convolutional neural network and adopt the residual network model to increase the model's fitting and stability. Secondly, the attention mechanism is integrated to reinforce the expression of each weight information, allowing the network to differentiate feature information more effectively and elevating the overall performance of the network. Thirdly, we use five sensors to form a local Internet of Things to collect a large amount of Qin opera audio data for experiments. Finally, multiple experiments confirm the effectiveness of the proposed model in the emotional classification of Qinqiang Opera.
What problem does this paper attempt to address?