Emotional Analysis and Personalized Recommendation Analysis in Music Performance

Abstract

Music performance belongs to music recreation activities, is through singing, musical instrument performance, and vocal conductor, including a variety of artistic means, to convey music to the audience with real sound effects that can be felt, and can play its social function. Emotional state is a part of the whole attitude, which is consistent with introverted feelings and intentions in attitude, and is a complex and stable physiological evaluation and experience of attitude in physiology. This paper first introduces the concepts of music performance from the perspective of music appreciation and music ability. Then based on the emotional feature learning of matrix factorization constrained nonnegative matrix factorization and model optimization algorithm are used to analyze the emotional aspects of music. Finally, through the experimental comparison of the difficulty of emotion in various artistic creations, a conclusion is drawn. The research involves matrix decomposition, mathematical modeling, model optimization, digital audio technology application, and other fields. In the third part of this paper, the constrained nonnegative matrix factorization, the constrained nonnegative matrix factorization with external information, and the model optimization algorithm are used to study the mathematical modeling.