Wavelet-Filtering of Symbolic Music Representations for Folk Tune Segmentation and Classification

The aim of this study is to evaluate a machine - learning method in which symbolic representations of folk songs are segmented and classified into tune families with Haar - wavelet filtering. The method is compared with previously proposed Gestalt - based method . Melodies are represented as discrete symbolic pitch - time signals. We apply the continuous wavelet transform (CWT) with the Haar wavelet at specific scales, obtaining fi l-tered versions of melodies emphasizing their information at pa r-ticular time - scales. W e use the filtered signal for representation and segmentation, using the wavelet coefficients' local maxima to indicate local boundaries and classify segments by means of k - nearest neighbours based on standard vector - metrics (Eucli dean, cityblock), and com pare the results to a Gestalt - based se g-mentation method and metrics applied directly to the pitch si g-nal. We found that the wavelet based segmentation and wavelet - filtering of the pitch signal lead to better classification accuracy in cross - validated evalu ation when the time - scale and other p a-rameters are optimized .