Optimization of Parameter Settings for Genetic Algorithms in Music Segmentation

Genetic algorithms have been introduced to the field of media segmentation including image, video, and also music segmentation since segmentation problems usually have complex fitness landscapes. Music segmentation can provide insight into the structure of a music composition so it is an important task in music information retrieval (MIR). The authors have already presented the application of genetic algorithms for the music segmentation problem in an earlier paper. This paper focuses on the optimization of parameter settings for genetic algorithms in the field of MIR as well as on the comparison of their results.