Repozytorium publikacji - Politechnika Gdańska

Music analysis and processing aims at understanding information retrieved from music (Music Information Retrieval). For the purpose of music data mining, machine learning (ML) methods or statistical approach are employed. Their primary task is recognition of musical instrument sounds, music genre or emotion contained in music, identification of audio, assessment of audio content, etc. In terms of computational approach, music databases contain imprecise, vague and indiscernible data objects. Moreover, most of the machine learning algorithms outcomes are given as a black-box result. Also, underfitting or overfitting may occur, meaning that either the model description is not complex enough or the test set is too small or not sufficiently representative. Thus the goal is to generalize the model. To overcome some of these problems, rule-based systems may be used, e.g., based on rough set theory that shows the outcome in the form of rules interconnecting features retrieved from music. Thus, first, principles of rule-based classifiers and particularly rough sets (RS) are presented, showing their usability in the music domain. A potential of the rough set-based approach was shown in the context of music genre recognition. The results were analyzed in terms of the recognition rate and computation time efficiency.

Autorzy

• Grazina Korvel,
• prof. dr hab. inż. Bożena Kostek link otwiera się w nowej karcie