Makine öğrenmesi ile müzik enstrüman sınıflandırması için bir çoklu-öznitelik yaklaşımı

Öz

Bu çalışma, RMS Energy, Zero Crossing Rate (ZCR) ve Spectral Centroid gibi spektral ses özniteliklerinin, Random Forest ve XGBoost algoritmaları kullanılarak müzik aleti sınıflandırmasındaki performansını incelemektedir. Bu makine öğrenmesi algoritmaları, karmaşık sınıflandırma senaryolarında artırılmış doğruluk sağlama ve yüksek korelasyonlu enstrüman sınıfları arasında ayrım yapabilme yeteneğine sahiptir. Bu çalışmada makine öğrenmesi yaklaşımları açıklanabilir ve hesaplama açısından verimli olmaları ve derin öğrenmenin donanım veya veri kısıtlamaları nedeniyle uygulanamadığı durumlarda kullanışlı olmaları sebebiyle tercih edilmiştir. Deneysel çalışmada, ses öznitelikleri 20 enstrüman sınıfını içeren Philharmonia veri setinden elde edilmiştir. Her öznitelik seti tek başına ve ikili ile üçlü kombinasyonları da dahil olmak üzere toplam yedi farklı konfigürasyon değerlendirilmiştir. Doğruluk açısından en yüksek performans, tüm öznitelikler bir arada kullanıldığında elde edilmiş olup Random Forest için 0,91 ve XGBoost için 0,93’tür. Bu makine öğrenmesi algoritmaları, müzikteki akustik farklılıkları ayırt etmede özellikle başarılı olmuştur. Karışıklık matrisi analizi, her iki modelin gitar ve banjo gibi belirgin akustik özelliklere sahip enstrümanlarda en iyi performansı gösterdiğini ortaya koymuştur. Bulgular, birden fazla tamamlayıcı özelliğin kombinasyonunun müzik aleti sınıflandırma performansını artırdığını göstermiştir.

Anahtar Kelimeler

• MIR
• enstrüman sınıflandırma
• makine öğrenmesi
• öznitelik çıkarımı

A multi-feature approach for musical instrument classification using machine learning

Öz

This study examines the performance of a collection of spectral audio features, including RMS Energy, Zero Crossing Rate (ZCR), and Spectral Centroid, for musical instrument classification by using the Random Forest and XGBoost classifiers. These machine learning algorithms demonstrate enhanced precision in complex classification scenarios and improve the ability to discriminate among highly correlated instrument classes. Machine learning approaches were employed in this study due to being explainable, computationally efficient, and suitable when deep learning is not feasible under the constraints of hardware or data. As part of the experimental setup, the audio features were obtained from the Philharmonia dataset, which includes 20 instrument classes. Seven different configurations were evaluated, including each feature set individually, as well as their pairwise and triplet combinations. The highest performance in terms of accuracy was obtained when all attributes were utilized: 0.91 with Random Forest and 0.93 with XGBoost. These machine learning algorithms were particularly well adapted to distinguish acoustic differences in music. Confusion matrix analysis indicated that both models worked best for instruments with clear acoustic characteristics, such as guitar and banjo. The findings suggested that the combination of multiple complementary features improves the classification performance of musical instruments.

Anahtar Kelimeler

• MIR
• instrument classification
• machine learning
• feature extraction

Kaynakça

1. J. McKay, Automatic musical instrument identification, Master’s thesis, Dublin Institute of Technology, 2011.
2. S. Murthy and S. G. Koolagudi, Content-based music information retrieval (cb-mir) and its applications toward the music industry: A review, ACM Computing Surveys (CSUR), vol. 51, no. 3, pp. 1–46, 2018.
3. C. Constantinescu and R. Brad, An overview of sound features in time and frequency domain, International Journal of Advanced Statistics and IT&C for Economics and Life Sciences, vol. 13, no. 1, 2023.
4. J. Pons, O. Nieto, M. Prockup, E. Schmidt, A. Ehmann, and X. Serra, End-to-end learning for music audio tagging at scale, arXiv preprint arXiv:1711.02520, 2017.
5. W. Qin and B. Yin, Environmental sound classification algorithm based on adaptive data padding, in 2022 International Seminar on Computer Science and Engineering Technology (SCSET), pp. 84–88, IEEE, 2022.
6. B. Toghiani-Rizi and M. Windmark, Musical instrument recognition using their distinctive characteristics in artificial neural networks, arXiv preprint arXiv:1705.04971, 2017.
7. P. Uruthiran and L. Ranathunga, Optimization of feature selection and classification of oriental music instruments identification, in 1st International Conference on Artificial Intelligence and Data Sciences (AiDAS), pp. 120–125, IEEE, 2019.
8. H. Tu and Y. Li, Neural network for music instrument identification, CS 229 Machine Learning Final Project, Stanford University, 2023.

9. Y. Su, Instrument classification using different machine learning and deep learning methods, Highlights in Science, Engineering and Technology, vol. 34, pp. 136–142, 2023.
10. R. Bhagyalakshmi and M. Anandaraju, Identification of specific musical instruments using machine learning models, Journal of Integrated Science and Technology, vol. 13, no. 5, pp. 1108–1108, 2025.
11. J. J. Bosch, J. Janer, F. Fuhrmann, and P. Herrera, A comparison of sound segregation techniques for predominant instrument recognition in musical audio signals, in 13th International Society for Music Information Retrieval Conference (ISMIR 2012), pp. 559–564, 2012.
12. R. M. Bittner, J. Salamon, M. Tierney, M. Mauch, C. Cannam, and J. P. Bello, MedleyDB: A multitrack dataset for annotation-intensive MIR research, in 15th International Society for Music Information Retrieval Conference (ISMIR 2014), vol. 14, pp. 155–160, 2014.
13. S. K. Mahanta, A. F. U. R. Khilji, and P. Pakray, Deep neural network for musical instrument recognition using MFCCs, Computación y Sistemas, vol. 25, no. 2, pp. 351–360, 2021.
14. G. M. Bhandari, Different audio feature extraction using segmentation, International Journal for Innovative Research in Science Technology, vol. 2, no. 9, pp. 1–5, 2016.
15. P. A. Babu, V. S. Nagaraju, and R. R. Vallabhuni, Speech emotion recognition system with “, in 10th IEEE International Conference on Communication Systems and Network Technologies (CSNT), pp. 421–424, IEEE, 2021.
16. D. Lavry, Sampling theory for digital audio, Lavry Engineering, Inc., 2004.
17. E. Bressert, SciPy and NumPy: An overview for developers, O’Reilly Media, Inc., 2012.
18. I. H. Sarker, Machine learning: Algorithms, real-world applications and research directions, SN Computer Science, vol. 2, no. 3, p. 160, 2021.
19. S. Fatima, A. Hussain, S. B. Amir, S. H. Ahmed, and S. M. H. Aslam, et al., XGBoost and random forest algorithms: An in-depth analysis, Pakistan Journal of Scientific Research, vol. 3, no. 1, pp. 26–31, 2023.
20. G. Kunapuli, Ensemble methods for machine learning, Simon and Schuster, 2023.
21. K. Friedrichs, N. Bauer, R. Martin, and C. Weihs, A computational study of auditory models in music recognition tasks for normal-hearing and hearing-impaired listeners, EURASIP Journal on Audio, Speech, and Music Processing, vol. 2017, pp. 1–22, 2017.
22. P. Bonissone, J. M. Cadenas, M. C. Garrido, and R. A. Díaz-Valladares, A fuzzy random forest, International Journal of Approximate Reasoning, vol. 51, no. 7, pp. 729–747, 2010.
23. C. Bentéjac, A. Csörgő, and G. Martínez-Muñoz, A comparative analysis of gradient boosting algorithms, Artificial Intelligence Review, vol. 54, pp. 1937–1967, 2021.
24. D. Nielsen, Tree boosting with XGBoost – why does XGBoost win every machine learning competition?, Master’s thesis, NTNU, 2016.
25. A. Samat, E. Li, W. Wang, S. Liu, C. Lin, and J. Abuduwaili, Meta-XGBoost for hyperspectral image classification using extended MSER-guided morphological profiles, Remote Sensing, vol. 12, no. 12, p. 1973, 2020.
26. G. Agostini, M. Longari, and E. Pollastri, Musical instrument timbres classification with spectral features, EURASIP Journal on Advances in Signal Processing, vol. 2003, pp. 1–10, 2003.