Rengidil-Neveser and Ruhnevaz-Buselik pieces for Machine Learning Classification
Description
This data set contains Turkish music pieces written with SymbTr notation. There are 12 Rengidil, 44 Neveser compositions, 22 Ruhnevaz and 73 Bûselik compositions. In addition, there 16 features which are extracted from the given notes, such as arithmetic mean of all the notes in the piece given in Holderian commas or the highest note of the piece.
Files
Steps to reproduce
16 features were extracted from the notation. These are: • firstNote: The first note of the piece • lastNote: The last note of the piece • lowNote: The lowest note of the piece • highNote: The highest note of the piece • AEmean: Arithmetic mean of all the notes in the piece given in Holderian commas. • W_AEmean: Arithmetic mean of all the notes in a piece, where each note is weighted by its rhythmic note values. • AEstd: Standard deviation of the Holderian comma values for all notes in a piece. S • W_AEstd: Standard deviation of the Holderian comma values for all notes in a piece, where each note is weighted by its rhythmic note value. • modeA, modeB, modeC: The first, second and third repeated note in the piece, representing the most used perdes. • W_modeA, W_modeB, W_modeC: The first, second and third repeated note in the piece, where each note is weighted by its rhythmic note value • AEmedian: The median of the Arel-Ezgi-Uzdilek comma values for all notes in a piece. • W_AEmedian: The median of the Holderian comma values for all notes in a piece, where each note is weighted by its rhythmic note value. Those features are available for both non-transposed pieces and transposed Rengidil and Ruhnevaz pieces.. Tranpositions are made as follows: Rengidil pieces were transposed frrom F4 to G4, adding 9 commas to each note, so that the tonic is equal to Neveser. Ruhnevaz pieces were transposed from E4 to A4 by adding 22 commas to each note so that tonic is equal to Buselik.