Musical pattern extraction using genetic algorithms

Abstract

This paper describes a research work in which we study the possibility of applying genetic algorithms to the extraction of musical patterns in monophonic musical pieces. Each individual in the population represents a possible segmentation of the piece being analysed. The goal is to find a segmentation that allows the identification of the most significant patterns of the piece. In order to calculate an individual’s fitness, all its segments are compared among each other. The bigger the area occupied by similar segments the better the quality of the segmentation.