SavedAdd your journal

Journals / Türk Deprem Araştırma Dergisi / 2026 / Cilt: 8 Sayı: 1

Post-Earthquake Building Damage State Classification Using Machine Learning: The Case of İzmir

Makine Öğrenmesi Yöntemleri ile Deprem Sonrası Bina Hasar Sınıfı Kestirimi: İzmir Örneği

Derya Yalçın(Afet ve Acil Durum Yönetimi Başkanlığı)0000-0001-8518-1303
Bahadır Aktuğ(ANKARA ÜNİVERSİTESİ, MÜHENDİSLİK FAKÜLTESİ)0000-0002-7995-4477
Pages
92–103
Publication Date
2026-04-30
🏛️Bibliographic metadata and full text access for this article are provided via DergiPark.
DergiPark Record ↗
✨

AI Summary & Key Takeaways

AI Insight
⚡ TL;DR(Too Long; Didn't Read)

The results demonstrate that machine learning-based approaches provide an effective framework for capturing spatial patterns of post-earthquake building damage and supporting risk-informed decision-making processes.

📌 3 Key Takeaways

🎯
Objective & Hypothesis:This study aims to estimate post-earthquake building damage classes in the Bayraklı and Bornova districts of İzmir following the 30 October 2020 Samos Island (Seferihisar–İzmir) earthquake using machine learning methods.
📊
Key Findings & Data:The dataset consists of three damage classes (low, moderate, and high) derived from field-based damage assessments and exhibits a severe class imbalance.
💡
Conclusion & Impact:The results demonstrate that machine learning-based approaches provide an effective framework for capturing spatial patterns of post-earthquake building damage and supporting risk-informed decision-making processes.
🔬Design:Modeling & Simulation
📍Scope:2020 Verileri
⚙️Variables:Hasar kestirimi · Makine öğrenmesi · Sisam Adası (Seferihisar-İzmir) depremi

Abstract

This study aims to estimate post-earthquake building damage classes in the Bayraklı and Bornova districts of İzmir following the 30 October 2020 Samos Island (Seferihisar–İzmir) earthquake using machine learning methods. The dataset consists of three damage classes (low, moderate, and high) derived from field-based damage assessments and exhibits a severe class imbalance. Therefore, model performances were evaluated using class-sensitive metrics. Decision Tree, Random Forest, Artificial Neural Networks, Support Vector Machines, and Extreme Gradient Boosting (XGBoost) algorithms were comparatively analyzed. The outputs of the best-performing model were transferred to a Geographic Information Systems environment to generate building-scale spatial damage estimation maps. The results demonstrate that machine learning-based approaches provide an effective framework for capturing spatial patterns of post-earthquake building damage and supporting risk-informed decision-making processes.

Özet

Bu çalışmada, 30 Ekim 2020 Sisam Adası (Seferihisar-İzmir) depremi sonrasında Bayraklı ve Bornova ilçelerinde saha gözlemlerine dayalı olarak belirlenen bina hasar sınıflarının, makine öğrenmesi yöntemleriyle tahmin edilmesi amaçlanmıştır. Düşük, orta ve yüksek olmak üzere üç hasar sınıfından oluşan veri seti, belirgin sınıf dengesizliği içermektedir. Bu nedenle model performansları, dengesiz veri setlerine duyarlı ölçütler kullanılarak değerlendirilmiştir. Karar ağacı (DT), rastgele orman (RF), yapay sinir ağları (ANN), destek vektör makineleri (SVM) ve Aşırı Gradyan Artırma Algoritması (XGBoost), algoritmaları karşılaştırmalı olarak analiz edilmiştir. En iyi performansı sağlayan modelin çıktıları Coğrafi Bilgi Sistemleri ortamına aktarılmış ve mekânsal hasar kestirim haritası üretilmiştir. Elde edilen bulgular, makine öğrenmesi yöntemlerinin deprem sonrası bina hasarlarının mekânsal örüntülerini modellemede etkili bir araç sunduğunu göstermektedir.

🔖

Cite This Article

Yalçın, D., Aktuğ, B. (2026). Post-Earthquake Building Damage State Classification Using Machine Learning: The Case of İzmir. Türk Deprem Araştırma Dergisi, Cilt: 8 Sayı: 1, 92–103. https://doi.org/10.46464/tdad.1802507

📄 Full Text PDF Viewer

📄

PDF görüntüleyici hazırlanıyor...

References (26)

  1. AFAD, 2022. 30 Ekim 2020 İzmir Depremi Hasar Tespit Verileri AYDES, Resmi yazı ile temin edilen kurum verisi.
  2. AFAD, 2023. 06 Şubat 2023 Pazarcık-Elbistan Kahramanmaraş (Mw: 7.7– Mw: 7.6) Depremleri Raporu, T.C. İçişleri Bakanlığı Afet ve Acil Durum Yönetimi Başkanlığı, Erişim adresi: https://deprem.afad.gov.tr/assets/pdf/Kahramanmara%C5%9F%20Depremi%20%20Raporu_02.06.2023.pdf
  3. Aktuğ B., Tiryakioğlu İ., Sözbilir H., Özener H., Özkaymak Ç., Yiğit C.O., Solak H.İ., Eyübagil E.E., Gelin B., Tatar O., Softa M., 2021. GPS derived finite source mechanism of the 30 October 2020 Samos earthquake Mw 6.9 in the Aegean extensional region, Turkish Journal of Earth Sciences, 30(8), 718-737.
  4. Breiman L., 2001. Random forests, Machine Learning, 45(1), 5-32.
  5. Chawla N.V., Bowyer K.W., Hall L.O., Kegelmeyer W.P., 2002. SMOTE Synthetic minority over sampling technique, Journal of Artificial Intelligence Research, 16, 321-357.
  6. Chen T., Guestrin C., 2016. XGBoost A scalable tree boosting system, Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 785-794.
  7. Cortes C., Vapnik V., 1995. Support vector networks, Machine Learning, 20(3), 273-297.
  8. Çetin K.O., Ozan C., Koyuncu A., 2021. 30 Ekim 2020 İzmir Depremi’nin jeoteknik değerlendirmesi, Jeoloji Mühendisliği Dergisi, 45(2), 115-132.
  9. Erdik M., 2013. Earthquake risk in Turkey, Seismological Research Letters, 84(6), 1007-1019.
  10. FEMA, 2012. Multi hazard loss estimation methodology earthquake model HAZUS MH technical manual, Federal Emergency Management Agency, Washington D.C., 712 p.
  11. Fernandez A., Garcia S., Galar M., Prati R.C., Krawczyk B., Herrera F., 2018. Learning from imbalanced data sets, Springer International Publishing, Cham, Switzerland, 347 p.
  12. He H., Garcia E.A., 2009. Learning from imbalanced data, IEEE Transactions on Knowledge and Data Engineering, 21(9), 1263-1284.
  13. İzmir Büyükşehir Belediyesi, 2024. Bina Kimlik Belgesi Veri Seti, Erişim adresi: https://acikveri.bizizmir.com/dataset/bina-kimlik-belgesi
  14. Joyner W.B., Boore D.M., 1981. Peak horizontal acceleration and velocity from strong-motion records including records from the 1979 Imperial Valley, California, earthquake. Bulletin of the Seismological Society of America, 71:2011-2038.
  15. Kappos A.J., Panagopoulos G., Panagiotopoulos C., 2006. A hybrid method for the vulnerability assessment of R C and URM buildings, Bulletin of Earthquake Engineering, 4, 391-413.
  16. Khosravani H., Rezaeian A., Motamedi M., 2021. Machine learning based vulnerability assessment of buildings under seismic excitation, Engineering Structures, 244, 112821.
  17. Onat E., İlki A., Yakut A., 2022. Earthquake damage prediction of reinforced concrete buildings using machine learning algorithms, Bulletin of Earthquake Engineering, 20(12), 12345-12367.
  18. Quinlan J.R., 1993. C4.5 Programs for machine learning, Morgan Kaufmann Publishers, San Mateo, CA, USA, 302 p.
  19. Rumelhart D.E., Hinton G.E., Williams R.J., 1986. Learning representations by back propagating errors, Nature, 323(6088), 533-536.
  20. Sengör A.M.C., Tüysüz O., İmren C., Sakınç M., Eyidoğan H., Görür N., Le Pichon X., Rangin C., 2014. The North Anatolian Fault A new look, Annual Review of Earth and Planetary Sciences, 32, 37-64.
  21. Seyitoğlu G., Aktuğ B., Esat K., Kaypak B., 2022. Neotectonics of Turkey and surrounding regions A new perspective with block modelling, Geologica Acta, 20, 1-21.
  22. Sun Y., Wong A.K.C., Kamel M.S., 2009. Classification of imbalanced data A review, International Journal of Pattern Recognition and Artificial Intelligence, 23(4), 687-719.
  23. Xie L., Zhu L., Chen Y., 2020. Machine learning techniques for seismic vulnerability assessment A review, Earthquake Engineering and Engineering Vibration, 19(3), 611-632.
  24. Yılmaz M.T., Akgün H., Erberik A., Gündoğan A.A., Koçkar M., Temiz A., Okalp K., Jalehforouzan A., Şahin G., 2023. Topoğrafik ve Jeolojik Sayısal Veriler Kullanılarak Türkiye Ölçeğinde Tahmini Vs30 Haritasının Cbs Ortamında Üretilmesi, Ulusal Deprem Araştırma Programı (UDAP-Ç-20-01) Proje Sonuç Raporu, Kasım 2023 Ankara, 97s.
  25. Yılmaz Z., Yıldırım H., 2022. Assessment of seismic vulnerability using random forest and GIS integration A case study for western Türkiye, Natural Hazards, 112(2), 1147-1168.
  26. Zhu L., Wang W., Li H., 2020. Seismic damage prediction of buildings based on artificial neural networks, Soil Dynamics and Earthquake Engineering, 136, 106206.