SIR Physics Investigation Abstract
COMPUTATIONAL PREDICTION OF SONG GENRE WITH RECURRENCE QUANTIFICATION ANALYSIS
Presenter:
Andrew Keller, Illinois Mathematics and Science Academy, 1500 West Sullivan Road, Aurora, IL 60506
Mentor:
Dr. Charles Webber, Loyola University Medical Center
Abstract:
Music has been a part of human culture for several millennia and yet there remain plenty of unanswered questions. Through the use of a tool called recurrence quantification analysis, I have investigated the possibility of whether or not there is a way to quantitatively classify songs into their respective genres. Such a task has been undertaken before with different means of analysis (e.g. Fourier transforms), but remains in the experimental stages. Though time consuming, this idea of classifying songs by their recurrence patterning seems to hold great promise as some preliminary results have already been obtained. The researcher believes that the results at least show a way of classifying a given song with high accuracy as rock, classical, experimental electronic, or otherwise. If the accuracy of prediction and speed of analysis improve, the results will prove promising for other aspects of signal processing, such as voice recognition.