Mining Contrast Patterns in Jazz Solos
摘要
Contrast pattern mining is the process of finding patterns whose frequency differs significantly from one group of data to another. In this study we apply a contrast pattern discovery algorithm to find sequential patterns distinctive within the solos of Miles Davis. A dataset containing transcriptions from his recordings from the 1950s and 1960s was mined in an effort to determine whether the patterns he employed in the earlier period remained constant in his later period, or whether new patterns emerged.