SIATEC-C: Computationally efficient repeated pattern discovery in polyphonic music

The use of point-set representations of music enable repeated pattern discoveryto be performed on polyphonic music. The discovery of patterns containing polyphony is also enabled by the use of point-set representations. The SIA and SIATEC algorithms discover repeated patterns in point-sets bycomputing maximal translatable patterns and their translational equivalence classes.While the algorithms are relatively efficient, their application to larger piecesof music is not viable due to quadratic space complexity.This paper introcudes a novel algorithm, SIATEC-C, for repeated pattern discovery in point-set representations of music. The algorithm discovers repeated patterns and finds all of their occurrences, whilerunning with subquadratic space complexity. The algorithm can also provide significant runningtime improvements over the comparable SIATEC algorithm.The computational performance of the algorithm is compared with SIATEC. The accuracy of the algorithmis also evaluated on the JKU-PDD data set.