Evaluation Method of Music Teaching Effect Based on Fusion of Deep Neural Network under the Background of Big Data

Abstract

Music is an art course that is different from traditional subjects. It not only needs to teach basic music knowledge but also needs to show the timbre contained in music itself, the form of music creation, or the effects expressed by music. The teaching effect evaluation model for traditional subjects can no longer be applied to the teaching effect evaluation of music subjects. If the traditional subject teaching effect evaluation model is used to evaluate the music subject teaching effect, it will easily lead to the inaccuracy of the effect evaluation. Because there is a strong subjectivity in the evaluation of the teaching effect of music subjects, this study combines deep neural network technology to study the feasibility and accuracy of the convolutional neural network (CNN) and long short-term memory neural network (LSTM) method in the evaluation of music teaching effects. It mainly reflects the evaluation under the three characteristics of music basic knowledge, music expression effect, and music innovation effect. The research results show that the CNN and LSTM methods can better extract the three features in the evaluation of music teaching effect, and the maximum prediction error is only 2.23%. Hybrid CNN-LSTM with LSTM neural network has higher accuracy in predicting music teaching effect than single neural network technique. The linear correlation coefficients also all exceeded 0.955.