Walker, Marcus2019-12-112019-12-112019-03https://hdl.handle.net/11299/208967University of Minnesota M.S. thesis. Marcus 2019. Major: Mathematics. Advisor: Marshall Hampton. 1 computer file (PDF); vi, 94 pages + 2 supplementary files.Music has powerful and inscrutable effects on the human mind, and we are far from fully understanding how that magic works. But music is not random: there are patterns in the sounds and rhythms of a piece that can be analyzed, things that can be learned! In this work I will review relevant research on the subject of Music Information Retrieval and then introduce Composobot, an original program that incorporates and extends the lessons of that research. Together we will examine how Composobot prepares musical pieces for processing, analyzes them to extract systems of patterns and dependencies, and then composes novel musical pieces based on what it has learned. Finally, we will discuss how much of the magic that is in the music we love can be captured by learning patterns the way Composobot does, and how those methods might be tweaked to capture an even greater share of it.enAlgorithmic CompositionMarkovMusic Information RetrievalThesis or Dissertation