Using Data Mining Techniques to Explore the Compositional Characteristics of Mozart’s Piano Concertos

Abstract

The piano concerto is the most important area of Mozart’s musical creation, except opera and throughout Mozart’s life. In this project, the original audio files of Mozart’s piano concertos were collected and preprocessed using data mining techniques. Then, the fast Fourier transform algorithm was used to extract the audio features, which was combined with the multimodal music emotion classification of audio and lyrics, to complete the fusion of multiple different music features. Finally, based on Thayer’s emotion model, the support vector machine algorithm is used to effectively classify music emotion and music frequency to deeply explore the creative characteristics of Mozart’s piano concerto. The results show that the algorithm in this paper identifies 52 samples that are biased towards calm emotion, which is 2.5 times more than the samples that are biased towards happy emotion, indicating that the style of Mozart’s music creation is biased towards calmness and elegance, and 80% of the Mozart’s music samples are biased towards high frequency above 4000 HZ, indicating that Mozart mainly creates high-frequency music. The algorithm in this paper accurately accomplishes the task of exploring the characteristics of Mozart’s piano concerto composition.