Characterizing and classifying music genres and subgenres via association analysis
Lefaivre, Adam T.
University of Lethbridge. Faculty of Arts and Science
Lethbridge, Alta. : Universtiy of Lethbridge, Department of Mathematics and Computer Science
In this thesis, we investigate the problem of automatic music genre classification in the field of Music Information Retrieval (MIR). MIR seeks to apply convenient automated solutions to many music-related tasks that are too tedious to perform by hand. These tasks often deal with vast quantities of music data. An effective automatic music genre classification approach may be useful for other tasks in MIR as well. Association analysis is a technique used to explore the inherent relationships among data objects in a problem domain. We present two novel approaches which capture genre characteristics through the use of association analysis on large music datasets. The first approach extracts the characteristic features of genres and uses these features to perform classification. The second approach attempts to improve on the first one by utilizing a pairwise dichotomy-like strategy. We then consider applying the second approach to the problem of automatic subgenre classification.
Music Information Retrieval , Genre Classification , Association Analysis , Data Mining , Machine Learning , Music -- Data processing , Musical analysis -- Data processing , Popular music genres -- Data processing , Association rule mining , Information retrieval