Tek Kanallı Akciğer Seslerinde Süzgeç Tipi Özellik Seçim Yöntemlerini Kullanarak Solunum Patolojisinin Teşhisi

Year 2022, Issue: 34, 374 - 380, 31.03.2022

Mustafa Alptekin Engin , Latif Akçay , Selim Aras

https://doi.org/10.31590/ejosat.1082560

Abstract

Yapılan çalışmada, tek kanallı yaygın akciğer sesleri kullanılarak patolojik ve sağlıklı denekler üzerinde detaylı bir öznitelik analizi gerçekleştirilmiştir. 94 kişiden elde edilen normal, ronküs, ince ral ve kaba ral seslerine ait 594 adet solunum döngüsünün otomatik tespiti ile elde edilen veri tabanı kullanılmıştır. Daha sonra, sıfır geçiş oranı, enerji, enerjinin entropisi, spektral merkezilik, bir çerçevenin yayılımı, spektral entropi, spektral akı, spektral devrilme, Mel-frekans cepstral katsayıları, harmonik oran, pencerenin temel frekansı ve renk vektörü öznitelik çıkarma yöntemleri veri tabanına uygulanmıştır. Sonsuz gizli öznitelik seçimi, sonsuz öznitelik seçimi, özvektör merkeziliği, minimum artıklık, maksimum ilgililik, relief, karşılıklı bilgi, laplace skoru, çoklu küme, fisher, denetimsiz ayrımcı, yerel öğrenmeye dayalı kümeleme, korelasyona dayalı öznitelik seçim yöntemleri eğitim aşamasında kullanılmıştır. Sınıflandırma için destek vektör makinesi, k en yakın komşu, naive bayes ve karar ağaçları algoritmaları kullanılmıştır. Sonuç olarak, öznitelik sayısı sınırlı olmadığı durumda, k en yakın komşuluk sınıflandırıcısı ve çoklu küme öznitelik seçim yöntemi kullanılarak %97,5 sınıflandırma doğruluğu elde edilmiştir. Öznitelik sayısı 3 ile sınırlandırıldığında ise k en yakın komşu sınıflandırıcısı ve özvektör merkeziliği veya sonsuz öznitelik seçimi yöntemleri kullanılarak %91,6 sınıflandırma doğruluğu elde edilmektedir.

Keywords

Akciğer sesleri, Solunum Döngüsü, Sınıflandırma, Öznitelik Çıkarımı, Öznitelik Seçimi.

Diagnosing Respiratory Pathology Using Filter-Type Feature Selection Methods in Single-Channel Lung Sounds

Abstract

In this study, detailed feature analysis was performed on the pathological and healthy subjects via single-channel common lung sounds. The database obtained from the automatic detection of 594 respiratory cycles of normal, rhonchi, fine crackle, and coarse crackle sounds obtained from 94 people was used. Then, zero-crossing rate, energy, the entropy of energy, spectral centroid, the spread of a frame, spectral entropy, spectral flux, spectral roll-off, Mel-frequency cepstral coefficients, harmonic ratio, the fundamental frequency of a window, and chroma vector feature extraction methods have been applied to the database. Infinite latent feature selection, infinite feature selection, eigenvector centrality, minimum redundancy maximum relevance, relief, mutual information, laplacian score, multi-cluster, fisher, unsupervised discriminative, local learning-based clustering, correlation-based feature selection methods are used in the training phase. Support vector machine, k nearest neighbors, naive bayes, and decision trees algorithms were used for classification. As a result, 97.5% classification accuracy was achieved by using the k nearest neighbor classifier and the multi-cluster feature selection method when the number of features is not limited. When the number of features is limited to 3, 91.6% classification accuracy is achieved by using the k nearest neighbor classifier and the eigenvector centrality or infinite feature selection methods.

Keywords

Lung Sounds, Respiratory Cycle, Classification, Feature Extraction, Feature Selection.

References

• Aras, S., Öztürk, M., & Gangal, A. (2018). Automatic detection of the respiratory cycle from recorded, single-channel sounds from lungs, ” Turk. Turk. J. Of Electr. Eng. Comput. SCI, 26, 11–22.
• Bartsch, M. A., & Wakefield, G. H. (2005). Audio thumbnailing of popular music using chroma based representations. IEEE Transactions on Multimedia, 7, 96–104.
• Berrar, D. (2018). Bayes’ theorem and naive Bayes classifier. Encyclopedia of Bioinformatics and Computational Biology: ABC of Bioinformatics. Amsterdam, The Netherlands: Elsevier Science Publisher.
• Bohadana, A., Izbicki, G., & Kraman, S. S. (2014). Fundamentals of lung auscultation, ” N. N. Engl. J. Med, 370(21).
• Cai, D., Zhang, C., & He, X. (2010). Unsupervised feature selection for multi-cluster data. Proceedings of the 16th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining - KDD ’10. New York, New York, USA: ACM Press.
• Ding, C., & Peng, H. C. (1989). Minimum Redundancy Feature Selection from Microarray Gene Expression Data. In Proc. Second IEEE Computational Systems Bioinformatics Conf (pp. 523–528). Boston: Kluwer Academic Publishers.
• Emeksiz, Z., & Bostancı, İ. (2018). Güncel Pediatri, c. 16, s. 3, ss. Akciğer: Sesimi Duyan Var Mı?, 79–84.
• Fix, E., & Hodges, J. L. (1951). Discriminatory analysis, nonparametric discrimination: consistency properties. USAF School of Aviation Medicine, 4.
• Forman, G. (2003). An Extensive Empirical Study of Feature Selection Metrics for Text Classification. Journal of Machine Learning Research, 3, 1289–1305.
• Gennari, J. H., Langley, P., & Fisher, D. (1989). Models of incremental concept formation, ” Artif. Artif. Intell, 40(1–3), 11–61.
• Göğüş, F. Z., Karlık, B., & Harman, G. (2016). Identification of pulmonary disorders by using different spectral analysis methods. International Journal of Computational Intelligence Systems, 9(4), 595. doi:10.1080/18756891.2016.1204110
• Gu, Q., Li, Z., & Han, J. (2012). Generalized Fisher score for feature selection. Retrieved from http://arxiv.org/abs/1202.3725
• Gurung, A., Scrafford, C. G., Tielsch, J. M., Levine, O. S., & Checkley, W. (2011). Computerized lung sound analysis as diagnostic aid for the detection of abnormal lung sounds: a systematic review and meta-analysis. Respiratory Medicine, 105(9), 1396–1403. doi:10.1016/j.rmed.2011.05.007
• Guyon, I., & Elisseeff, A. (2003). An Introduction to Variable and Feature Selection. JMLR, 3, 1157–1182.
• Guyon, I., Gunn, S., Nikravesh, M., & Zadeh, L. A. (2006). Feature Extraction: Foundations and Applications. Berlin, Germany: Springer.
• Hall, M. A. (1999). Correlation-based Feature Selection for Machine Learning. Hamilton.
• Kim, H. G., Moreau, N., & Sikora, T. (2007). MPEG-7 audio and beyond: Audio content indexing and retrieval. Nashville, TN: John Wiley & Sons.
• Kim, Y., Hyon, Y., & Jung, S. S. (2021). Respiratory sound classification for crackles, wheezes, and rhonchi in the clinical field using deep learning”. Sci Rep, 11.
• Koeipensri, T., Boonchoo, P., & Sueaseenak, D. (2016). The development of biosignal processing system (BPS-SWU V1. 0) for learning and research in biomedical engineering. In 9th Biomedical Engineering International Conference (BMEiCON), Laung Prabang (pp. 1–4). Laos.
• Ladha, L., & Deepa, T. (2011). Feature Selection Methods And Algorithms. International Journal on Computer Science and Engineering, 3(5), 1787–1797.
• Lehrer, S. (2018). Understanding lung sounds: Third edition. North Charleston, SC: Createspace Independent Publishing Platform.
• Liu, H., & Motoda, H. (2007). Computational methods of feature selection. London, England: CRC Press.
• Metlek, S., & Kayaalp, K. (2020). Makine Öğrenmesinde Teoriden Örnek Matlab Uygulamalarına Kadar Destek Vektör Makineleri. Ankara, Türkiye: İktisad Yayınları.
• Özkaya, U., Öztürk, Ş., & Barstugan, M. (2020). Coronavirus (COVID-19) classification using deep features fusion and ranking technique. In Big Data Analytics and Artificial Intelligence Against COVID-19: Innovation Vision and Approach (pp. 281-295). Springer, Cham.
• Palaniappan, R., Sundaraj, K., & Lam, C. K. (2016). Reliable system for respiratory pathology classification from breath sound signals. 2016 International Conference on System Reliability and Science (ICSRS). IEEE.
• Roffo, G., & Melzi, S. (2016). Features selection via eigenvector centrality. In Proceedings of New Frontiers in Mining Complex Patterns.
• Roffo, Giorgio, Melzi, S., Castellani, U., & Vinciarelli, A. (2017). Infinite latent feature selection: A probabilistic latent graph-based ranking approach. 2017 IEEE International Conference on Computer Vision (ICCV). IEEE.
• Roffo, Giorgio, Melzi, S., Castellani, U., Vinciarelli, A., & Cristani, M. (2021). Infinite Feature Selection: A graph-based feature filtering approach. IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(12), 4396–4410. doi:10.1109/TPAMI.2020.3002843
• Sankur, B., Kahya, Y. P., Çağatay Güler, E., & Engin, T. (1994). Comparison of AR-based algorithms for respiratory sounds classification. Computers in Biology and Medicine, 24(1), 67–76. doi:10.1016/0010-4825(94)90038-8
• Şen, I., Saraclar, M., & Kahya, Y. P. (2015). A Comparison of DVM and GMM-Based Classifier Configurations for Diagnostic Classification of Pulmonary Sounds”. IEEE Transactions on Biomedical Engineering, 62(7), 1768–1776.
• Sezer, E. A., Bozkır, A. S., Yağız, S., & Gökçeoğlu, C. (2010). Karar ağacı derinliğinin CART algoritmasında kestirim kapasitesine etkisi: bir tünel açma makinesinin ilerleme hızı üzerinde uygulama. Akıllı Sistemlerde Yenilikler ve Uygulamaları Sempozyumu.
• Yang, X.-K., He, L., Qu, D., Zhang, W.-Q., & Johnson, M. T. (2016). Semi-supervised feature selection for audio classification based on constraint compensated Laplacian score. EURASIP Journal on Audio, Speech, and Music Processing, 2016(1). doi:10.1186/s13636-016-0086-9
• Yang, Y., Shen, H. T., Ma, Z., & Et, A. (2011). L2,1-norm regularized discriminative feature selection for unsupervised learning. In Conf. International Joint Conference on Artificial Intelligence (pp. 1589–1594).
• Yilmaz, C. A., & Kahya, Y. P. (2006). Multi-channel classification of respiratory sounds”. In 28th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) (pp. 2864–2867). New York, USA.
• Zaffalon, M., & Hutter, M. (2002). Robust feature selection using distributions of mutual information. In Proceedings of the 18th International Conference on Uncertainty in Artificial Intellegence (UAI-2002) (pp. 577–584). San Francisco, CA.
• Zeng, H., & Cheung, Y.-M. (2011). Feature selection and kernel learning for Local Learning-Based Clustering. IEEE Transactions on Pattern Analysis and Machine Intelligence, 33(8), 1532–1547. doi:10.1109/TPAMI.2010.215