Pattern Detection and Discovery: The Case of Music Data Mining

摘要

In this paper the problem of automatically detecting (or extracting, inducing, discovering) patterns from music data, is addressed. More specifically, approaches for extracting "sequential patterns" from sequences of notes (and rests) are presented and commented. Peculiarities of music data have direct impact on the very nature of pattern extraction and, correlatively, on approaches and algorithms for carrying it out. This impact is analyzed and paralleled with other kinds of data. Applications of musical pattern detection are covered, ranging from music analysis to music information retrieval.