Perdana, Ramadhan Rosihadi (2017) Implementasi Ekstraksi Fitur Untuk Pengelompokan Berkas Musik Berdasarkan Kemiripan Karakteristik. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
5112100032-Undergraduate_Theses .pdf - Published Version Download (2MB) | Preview |
Abstract
Pengelompokkan musik berdasarkan karakteristik suara merupakan hal penting bagi penikmat musik.. Penikmat musik tidaklah mencari musik berdasarkan artis tetapi juga mencari musik berdasarkan genre yang diinginkannya. Karena itu dibutuhkan metode pengelompokkan yang tepat untuk mengkategorikan musik berdasarkan genre secara otomatis.
Tugas akhir ini melakukan pengelompokkan musik berdasarkan genre. Dengan mengekstraksi fitur spectral centroid, spectral flux, spectral rolloff, dan short time energy pada tiap berkas musik yang diolah dan kemudian dihitung nilai mean, median, skewness, dan kurtosisnya. Dan selanjutnya dikelompokkan menggunakan metode klasifikasi Random Forest dengan alat bantu Weka.
Uji coba dilakukan dengan menggunakan kombinasi nilai atribut komponen ekstraksi fitur dan berkas musik yang berbeda-beda sesuai genre. Hasil uji coba klasifikasi pada Tugas Akhir ini menghasilkan nilai akurasi terbaik sebesar 80,47%.
===========================================================================================================================================================================Music classification based on voice characteristics is essential for music lovers. Connoisseurs of music is not looking for music just by artists but also search for music by genre he/she wants. Therefore we need a method of grouping the right to categorize music by genre automatically.
This final task is grouping music by genre. By extracting feature spectral centroid, spectral flux, spectral rolloff, spectral flatness, zero crossing, and short-time energy.Of each music file is processed and then calculated the mean, median, skewness, and kurtosis. And further classified using Random Forest classification methods with Weka .
The test is done using a combination of the component of features extraction proccess and music files based on genre. The classification accuracy of 80,47% is achieved in this work.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Uncontrolled Keywords: | Ekstraksi Fitur Audio, Musik, Klasifikasi. |
| Subjects: | M Music and Books on Music > ML Literature of music Q Science > QA Mathematics T Technology > T Technology (General) |
| Divisions: | Faculty of Information Technology > Informatics Engineering > 55201-(S1) Undergraduate Thesis |
| Depositing User: | RAMADHAN ROSIHADI PERDANA |
| Date Deposited: | 04 May 2017 02:15 |
| Last Modified: | 08 Mar 2019 06:15 |
| URI: | http://repository.its.ac.id/id/eprint/3599 |
Actions (login required)
 |
View Item |