Mobile Platform for MOCC Music Hybrid Teaching Based on Convolutional Neural Network

Abstract

With the wide application of Internet mobile devices in many industries, the hybrid teaching mode of online and offline has also become a research hotspot in the field of education. In the process of constructing the music online teaching platform based on mobile platform MOOC, the distribution of samples will affect the data recognition results of the system. Therefore, this paper uses convolutional neural network as the backbone network to extract data features and improve the resolution. At the same time, in the process of data compression, this paper realizes the global average pool and unified parameters by improving the attention mechanism, so that all parameters interact with their K adjacent parameter characteristics, so as to reduce the overdependence between channels and reduce the complexity of the overall calculation. For the analysis of complexity, this paper detects the time required for serial training and parallel training, intercepts the average value of the parameters of all nodes, and obtains the optimal number of nodes of this model. Finally, combined with the characteristics of music teaching, this paper designs a mobile MOOC music teaching platform based on convolutional neural network. The platform includes modules such as basic information management and music course resource construction and applies it to the actual music course teaching process. The performance of the algorithm and the feasibility of the system are verified by classroom activity test, hoping to provide some reference for the research in the field of music mixed education.