Research work · 2017
Discovering motifs with variants in music databases
Benamar, Riyadh, Christine Largeron, Véronique Églin, Mylène Pardoen
HAL
Abstract
Music score analysis is an ongoing issue for musicologists.Discovering frequent musical motifs with variants is needed in order tomake critical study of music scores and investigate compositions styles.We introduce a mining algorithm, called CSMA for Constrained StringMining Algorithm), to meet this need considering symbol-based representationof music scores. This algorithm, through motif length and maximalgap constraints, is able to find identical motifs present in a singlestring or a set of strings. It is embedded into a complete data miningprocess aiming at finding variants of musical motif. Experiments, carriedout on several datasets, showed that CSMA is efficient as string miningalgorithm applied on one string or a set of strings.