Building collapse prediction due to earthquakes through ML algorithm: Hybrid methodology based on past-earthquake damage reports and resilience index

Abstract

This study presents a hybrid methodology for predicting building collapses within the Intelligent Circular Resilience ( ICR) framework. This uses a supervised Machine Learning ( ML) approach, earthquake damage reports, and the Simplified Resilience Index (SRI), derived from existing earthquake damage models (EDM)-based on fragility and vulnerability functions-used in the probabilistic seismic risk assessment ( PSRA). A curated building damage database comprising 89 structures (71 collapsed and 18 non-collapsed) from ten countries affected by major earthquakes (Mw 6.1–8.1, epicentral distances of 3–125 km, and PGA values ranging from 0.14 g to 0.82 g) was developed, including attributes related to exposure: occupancy, main structural material, number of stories, construction year, and hazard: magnitude, epicentral distance, intensity measures (Peak-ground acceleration, PGA, and elastic spectral acceleration). The dataset includes events such as the 2017 Puebla–Morelos earthquake (Mw 7.1, Mexico), the 1999 Kocaeli earthquake (Mw 7.6, Turkey), and the 2011 Christchurch earthquake (Mw 6.1, New Zealand), among others. Likewise, dependent attributes such as time elapsed and SRI (under 120-, 180-, and 365-day recovery scenarios) were calculated using 2-EDMs. Eight Random Forest models were trained and tested for collapse and non-collapse classification using combinations of independent and dependent attributes. The results indicate that models incorporating exposure-related variables-such as structural material, number of stories, construction year, and occupancy-alongside the SRI significantly improve collapse classification performance, achieving recall and F1 scores above 95%. Notably, many collapsed buildings exhibited low intensities (PGA ≤ 0.25 g), emphasizing the influence of local site effects-particularly in Mexico City. The findings demonstrate that incorporating SRI enhances the reliability of collapse prediction and supports its use as an interpretable resilience proxy during early ICR stages. This hybrid methodology bridges empirical data, traditional PSRA models, and ML techniques, contributing to more accurate and scalable post-earthquake resilience assessments.

Keywords

Seismic resilience / Simplified resilience index / Machine learning / Fragility and vulnerability functions / Collapse prediction

References

[1] Bradley BA. Strong ground motion characteristics observed in the 13 June 2011 Mw6.0 Christchurch, New Zealand earthquake. Soil Dyn Earthq Eng 2010; 91: 23-38. https://doi.org/10.1016/j.soildyn.2016.09.006. 2016.

[2] Tena-Colunga A. The recovery process of housing in Mexico City 7+ years after the 2017 Puebla-Morelos earthquake. Resilient Cities Struct 2025; 4(3): 67-98. https://doi.org/10.1016/j.rcns.2025.08.002.

[3] Alberto Y, Otsubo M, Kyokawa H, Kiyota T, Towhata I. Reconnaissance of the 2017 Puebla, Mexico earthquake. Soils Found 2018; 58(5): 1073-92. https://doi.org/10.1016/j.sandf.2018.06.007.

[4] Dinçer AE, Dincer NN, Tekin-Koru A, Yaşar B, Yılmaz Z. The impact of Kahramanmaras¸ (2023) earthquakes: a comparative case study for Adıyaman and Malatya. Int J Disaster Risk Reduct 2024; 110: 104647. https://doi.org/10.1016/j.ijdrr.2024.104647. February.

[5] Tonyalı İ, Akbas SO, Beyaz T, Kayabalı K, Gokceoglu C. Case study of a foundation failure induced by cyclic softening of clay during the 2023 Kahramanmaras¸ earthquakes. Eng Geol 2024; 332: 107477. https://doi.org/10.1016/j.enggeo.2024.107477.

[6] Karakas VE, Yildiz I, Can R, Kocaman S, Onder Cetin K, Gokceoglu C. Event-based regional model of liquefaction susceptibility in Amik, Gölbas¸ı and Kahramanmaras¸ Basins after the February 6, 2023 earthquakes. Eng Geol 2024; 339: 107644. https://doi.org/10.1016/j.enggeo.2024.107644.

[7] Ecemis N, Donmez C, Karaman M, Valizadeh H, Dalgic KD. Assessment of seismic liquefaction and structural instability in Adıyaman-Gölbas¸ı after the February 6, 2023, earthquakes in Türkiye. Soil Dyn Earthq Eng 2025; 192: 109300. https://doi.org/10.1016/j.soildyn.2025.109300.

[8] Karray M, et al. Türkiye Mw 7.7 Pazarcık and Mw 7.6 Elbistan earthquakes of February 6th, 2023: contribution of valley effects on damage pattern. Soil Dyn Earthq Eng 2024; 181: 108634. https://doi.org/10.1016/j.soildyn.2024.108634. April.

[9] Yıldız Ö, Çabalar AF. Analyzing seismic impact across regions: a comprehensive study of the Pazarcık and Elbistan earthquakes in Türkiye (February 6, 2023). J Afr Earth Sci 2025; 223. https://doi.org/10.1016/j.jafrearsci.2024.105463. August 2024.

[10] Yetkin M, Dedeoglu IO, Erkek H, Calayir Y. Damage/collapse types of minarets for the February 6, 2023 Pazarcık and Elbistan earthquakes: comparative evaluation of field studies and numerical analyses. Thin-Walled Struct 2026; 218: 114033. https://doi.org/10.1016/j.tws.2025.114033. no. PC.

[11] Arslan MH, Dere Y, Ecemiş AS, Doğan G, Özturk M, Korkmaz SZ. Code-based damage assessment of existing precast industrial buildings following the February 6th, 2023 Kahramanmaras¸ earthquakes (Pazarcık Mw 7.7 and Elbistan Mw7.6). J Build Eng 2024; 86: 108811. https://doi.org/10.1016/j.jobe.2024.108811. February.

[12] Şen F, Sunca F, Altunişik AC. Seismic performance assessment of a base-isolated hospital building subjected to February 6, 2023, Kahramanmaras¸, Türkiye earthquakes (Mw 7.7 Pazarcık and Mw 7.6 Elbistan) and seismic fragility analysis considering different construction stages. Soil Dyn Earthq Eng 2024; 185. https://doi.org/10.1016/j.soildyn.2024.108876. July.

[13] F. Avcil, A. Büyüksaraç, and R. Izol , “ Structural damages in masonry buildings in Adıyaman during the Kahramanmaras¸ (Turkiye) earthquakes (M w 7 . 7 and M w 7 . 6) on,” vol. 151, no. June, 2023.

[14] Gurbuz T, Cengiz A. Structural damages during the February 06, 2023 Kahramanmaras¸ Earthquakes in Turkey. Soil Dyn Earthq Eng 2024; 191: 109214. https://doi.org/10.1016/j.soildyn.2025.109214. 2025.

[15] Atmaca B, et al. What is the reason for collapses on February 6, 2023 Kahramanmaras¸ earthquakes: insights from a dataset of 400 collapsed RC buildings post-earthquake analysis. J Build Eng 2025; 107. https://doi.org/10.1016/j.jobe.2025.112660. April.

[16] Sivrikaya O, et al. Impact of soil conditions and seismic codes on collapsed structures during the 2023 Kahramanmaras¸ earthquakes: an in-depth study of 400 reinforced concrete buildings. Soil Dyn Earthq Eng 2024; 190: 2025. https://doi.org/10.1016/j.soildyn.2024.109119.

[17] Aroquipa H, Hurtado A, Angel C, Aroquipa A, Gamarra A, Almeida Del Savio A. A cost-benefit analysis for the appraisal of social and market prices in the probabilistic seismic risk assessment of building portfolios: a methodology for the evaluation of disaster risk reduction programs. Int J Disaster Risk Reduct 2022; 90: 103637. https://doi.org/10.1016/j.ijdrr.2023.103637. May 2023.

[18] Bruneau M, et al. A framework to quantitatively assess and enhance the seismic resilience of communities. Earthq Spectra 2003; 19(4): 733-52. https://doi.org/10.1193/1.1623497.

[19] Cimellaro PG, Reinhorn AM, Bruneau M. Framework for analytical quantification of disaster resilience. Eng Struct 2010; 32(11): 3639-49. https://doi.org/10.1016/j.engstruct.2010.08.008.

[20] Cutter SL, Burton CG, Emrich CT. Disaster resilience indicators for benchmarking baseline conditions. J Homel Secur Emerg Manag 2010; 7(1). https://doi.org/10.2202/1547-7355.1732.

[21] Argyroudis SA, Mitoulis SA, Hofer L, Zanini MA, Tubaldi E, Frangopol DM. Resilience assessment framework for critical infrastructure in a multi-hazard environment: case study on transport assets. Sci Total Environ 2020; 714: 136854. https://doi.org/10.1016/j.scitotenv.2020.136854.

[22] Salgado-Gálvez MA, et al. Development of a regionally consistent and fully probabilistic earthquake risk model for Central Asia. Nat Hazards Earth Syst Sci 2024; 24(11): 3851-68. https://doi.org/10.5194/nhess-24-3851-2024.

[23] Rudman A, Douglas J, Tubaldi E. The assessment of probabilistic seismic risk using ground-motion simulations via a Monte Carlo approach. Nat Hazards 2024; 120(7): 6833-52. https://doi.org/10.1007/s11069-024-06497-1.

[24] L. Pujades, Y.F. Vargas-alzate, and J.R. González-drigo , “ Un enfoque probabilista de la fragilidad y daño sísmico esperado en edificios porticados de hormigón armado Fragility and expected seismic damage in reinforced concrete framed buildings. A probabilistic approach,” vol. 7012, pp. 67-102, 2019.

[25] Trevlopoulos K, Gehl P, Negulescu C, Crowley H, Danciu L. Comparing components for seismic risk modelling using data from the 2019 Le Teil (France) earthquake. Nat Hazards Earth Syst Sci 2024; 24(7): 2383-401. https://doi.org/10.5194/nhess-24-2383-2024.

[26] Xin D, Zhang Z, Chen B, Wenzel F, Li Y, Chen X. Can we develop a more targeted approach to mitigating seismic risk? npj Nat Hazards 2024; 1(1): 19. https://doi.org/10.1038/s44304-024-00020-z.

[27] Al-Nammari F, Alzaghal M. Towards local disaster risk reduction in developing countries: challenges from Jordan. Int J Disaster Risk Reduct 2015; 12: 34-41. https://doi.org/10.1016/j.ijdrr.2014.11.005.

[28] Etinay N, Egbu C, Murray V. Building urban resilience for disaster risk management and disaster risk reduction. Procedia Eng 2017; 212: 575-82. https://doi.org/10.1016/j.proeng.2018.01.074. 2018.

[29] Xu J, Xing Z, Hu Y. Post-earthquake damage assessment of beam-to-column bolted and welded hybrid connections after fire based on machine learning. Structures 2025; 84: 110951. https://doi.org/10.1016/j.istruc.2025.110951. 2026.

[30] Belhadj FZ, Belhadj AF, Chabaat M. Machine-learning seismic damage assessment model for building structures. Artif Intell Geosci 2025; 6(2): 100155. https://doi.org/10.1016/j.aiig.2025.100155.

[31] Mostofi S, Yilmaz Z, Başağa HB, Okur FY, Altunişik AC, Taciroglu E. A hybrid stacked ensemble model for rapid seismic damage assessment with imbalanced training data: a case study on the 2023 Kahramanmaras¸ earthquakes. Eng Struct 2025; 340: 0-1. https://doi.org/10.1016/j.engstruct.2025.120754. May.

[32] Zorkun NG, Mutevelli Oncul MH, Karatay S, Erken F, Budak C. Identification of the ionospheric disturbances prior to large earthquakes in Türkiye using Random Forests algorithm. Adv Space Res 2025. https://doi.org/10.1016/j.asr.2025.06.080. xxxx.

[33] Mousavian P, Tafreshi ST, Majidian A, Di-Sarno L. Structural damage evaluation in RC buildings through ensemble learning: a comprehensive study of different techniques for efficient and reliable identification. Structures 2025; 76: 108831. https://doi.org/10.1016/j.istruc.2025.108831. April.

[34] Li L, Chang-Richards A, Boston M, Elwood K, Brunsdon D. A machine learning-supported rapid classification of building damage following the 2010-2011 Canterbury earthquakes. J Build Eng 2025; 107. https://doi.org/10.1016/j.jobe.2025.112807. no. April.

[35] Li Z, Lei H, Ma E, Lai J, Qiu J. Ensemble technique to predict post-earthquake damage of buildings integrating tree-based models and tabular neural networks. Comput Struct 2023; 287: 107114. https://doi.org/10.1016/j.compstruc.2023.107114.

[36] Hurtado A, Aroquipa H, Restrepo D, Nieto F, Carrillo J. Reliability of existing damage models based on past earthquake performance of buildings: a case study for collapse prediction using MDR-value and ML algorithm. Bull Earthq Eng 2025. BEEE-D-25-00416Under Review.

[37] Breiman L. Random forests. Mach Learn 2001; 45(1): 5-32. https://doi.org/10.1023/A:1010933404324.

[38] Mangalathu S, Sun H, Nweke CC, Yi Z, Burton HV. Classifying earthquake damage to buildings using machine learning. Earthq Spectra 2020; 36(1): 183-208. https://doi.org/10.1177/8755293019878137.

[39] Massone LM, Aceituno D, Carrillo J. Cumulative damage in RC frame buildings-the 2017 Mexico earthquake case. Adv Comput Des 2023; 8(1): 13-36. https://doi.org/10.12989/acd.2023.8.1.015.

[40] Galvis FA, Miranda E, Heresi P, Dávalos H, Ruiz-García J. Overview of collapsed buildings in Mexico City after the 19 September 2017 (M w 7.1) earthquake. Earthq Spectra 2020; 36(2_suppl): 83-109. https://doi.org/10.1177/8755293020936694.

[41] E.E.R.I.EERI , “ Concrete building Reports database-The concrete coalition,” A project of the Earthquake Engineering Research Institute. 2024. [Online]. Available: https://www.concretecoalition.org/projects/damaged-buildings-database/.

[42] Kircher CA, Whitman RV, Holmes WT. HAZUS earthquake loss estimation methods. Nat Hazards Rev 2006; 7(2): 45-59. https://doi.org/10.1061/(ASCE)1527-6988(2006)7:2(45).

[43] Whitman RV, Anagnos T, Kircher CA, Lagorio HJ, Lawson RS, Schneider P. Development of a national earthquake loss estimation methodology. Earthq Spectra 1997; 13(4): 643-61.

[44] Yamin LE, Hurtado AI, Barbat AH, Cardona OD. Seismic and wind vulnerability assessment for the GAR-13 global risk assessment. Int J Disaster Risk Reduct 2014; 10: 452-60. https://doi.org/10.1016/j.ijdrr.2014.05.007. PB.

[45] Adhikari RK, et al. GLOSI taxonomy: a tool for ‘seismic risk assessment’ oriented classification of school buildings. Int J Disaster Risk Reduct 2023; 87. https://doi.org/10.1016/j.ijdrr.2023.103594. February.

[46] Zhang Y, Burton HV, Sun H, Shokrabadi M. A machine learning framework for assessing post-earthquake structural safety. Struct Saf 2018; 72: 1-16. https://doi.org/10.1016/j.strusafe.2017.12.001.

[47] Zhou Z, Wang M, Han M, Yu X, Lu D. Prediction of damage potential in mainshock-aftershock sequences using machine learning algorithms. Earthq Eng Eng Vib 2024; 23(4): 919-38. https://doi.org/10.1007/s11803-024-2280-6.

[48] Kang S, et al. Deep learning method for post-earthquake damage assessment of frame structures based on time-Frequency analysis and CGAN. Earth Syst Environ 2024. https://doi.org/10.1007/s41748-024-00458-1.

[49] D. C., U. S. National Institute of Building Sciences (Washington, Earthquake loss estimation methodology: HAZUS technical manual . [Washington]: Federal Emergency Management Agency, 1997. [Online]. Available: https://catalog.hathitrust.org/Record/007822546.

[50] H. Pérez et al. , “ Reporte MNRS 005-2023: acuerdos metodológicos para la evaluación del riesgo sísmico en centros poblados,” Bogotá Colombia, 2023. [Online]. Available: https://www.sgc.gov.co/.

[51] Aroquipa H, Hurtado A. Seismic resilience assessment of buildings: a simplified methodological approach through conventional seismic risk assessment. Int J Disaster Risk Reduct 2022; 77: 103047. https://doi.org/10.1016/j.ijdrr.2022.103047. February.

[52] Aroquipa H, et al. Simplified methodological approach for estimating the mean repair time of building portfolios directed to the development of seismic resilience policies, based on the distribution of resources. J Build Pathol Rehabil 2023; 8(2): 72. https://doi.org/10.1007/s41024-023-00321-2.

[53] Li SQ, Liu HB. Statistical and vulnerability prediction model considering empirical seismic damage to masonry structures. Structures 2022; 39(74): 147-63. https://doi.org/10.1016/j.istruc.2022.03.024.

[54] Yamin LE, Hurtado A, Rincon R, Dorado JF, Reyes JC. Probabilistic seismic vulnerability assessment of buildings in terms of economic losses. Eng Struct 2017; 138: 308-23. https://doi.org/10.1016/j.engstruct.2017.02.013.

[55] Li S, Li C, Huang Y, Zhai C. Real-time seismic damage simulation for urban building portfolio based on basic building information and machine learning. Int J Disaster Risk Reduct 2024; 111: 104687. https://doi.org/10.1016/j.ijdrr.2024.104687. June.

[56] Wang W, Li L, Qu Z. Machine learning-based collapse prediction for post-earthquake damaged RC columns under subsequent earthquakes. Soil Dyn Earthq Eng 2023; 172: 108036. https://doi.org/10.1016/j.soildyn.2023.108036. February.

[57] Nguyen HD, Dao ND, Shin M. Capability of machine learning to predict seismic damage states of reinforced concrete wall structures. J Build Eng 2024; 106: 112620. https://doi.org/10.1016/j.jobe.2025.112620. 2025.

[58] Kazemi F, Asgarkhani N, Jankowski R. Machine learning-based seismic fragility and seismic vulnerability assessment of reinforced concrete structures. Soil Dyn Earthq Eng 2023; 166: 107761. https://doi.org/10.1016/j.soildyn.2023.107761.

[59] Bhatta S, Dang J. Seismic damage prediction of RC buildings using machine learning. Earthq Eng Struct Dyn 2023; 52(11): 3504-27. https://doi.org/10.1002/eqe.3907.

[60] Vasugi V, Santhi MH, Malathi G. Global seismic damage assessment of RC framed buildings using machine learning techniques. Artificial intelligence applications for sustainable construction. Elsevier; 2024. p. 389-404. https://doi.org/10.1016/B978-0-443-13191-2.00011-0.

[61] Li X, Mostafavi A. Machine learning approach for disaster risk and resilience assessment in coupled human infrastructure systems performance. npj Nat Hazards 2025; 2(1): 56. https://doi.org/10.1038/s44304-025-00104-4.

[62] G. Solomos and A. Caverzan , Review on resilience in literature and standards for critical built-infrastructure. 2014. doi: 10.2788/872668.

[63] Elkady S, Hernantes J, Labaka L. Decision-making for community resilience: a review of decision support systems and their applications. Heliyon 2024; 10(12): e33116. https://doi.org/10.1016/j.heliyon.2024.e33116.

[64] Suarez D, Gomez C, Medaglia AL, Akhavan-Tabatabaei R, Grajales S. Integrated decision support for Disaster risk management: aiding preparedness and response decisions in wildfire management. Inf Syst Res 2024; 35(2): 609-28. https://doi.org/10.1287/isre.2022.0118.

[65] Bellalem F, et al. Seismic risk assessment for the downtown of the city of Blida, Algeria. Int J Disaster Risk Reduct 2024; 103: 104314. https://doi.org/10.1016/j.ijdrr.2024.104314. February.

[66] Chee Yin H, Kassem MM, Mohamed Nazri F, El Maissi AM. Development of seismic resilience index methodology for RC structures via functionality curve. MethodsX 2022; 9: 101865. https://doi.org/10.1016/j.mex.2022.101865.

[67] Wu H, Guo E, Yan P, Liu J. A quantitative evaluation model for the seismic resilience of water supply systems based on fragility analysis. Sustainability 2023; 15(16): 12137. https://doi.org/10.3390/su151612137.

[68] Cha GW, Moon HJ, Kim YC. Comparison of random forest and gradient boosting machine models for predicting demolition waste based on small datasets and categorical variables. Int J Environ Res Public Health 2021; 18(16): 8530. https://doi.org/10.3390/ijerph18168530.

[69] Al-Sahaf H, et al. A survey on evolutionary machine learning. J R Soc NZ 2019; 49(2): 205-28. https://doi.org/10.1080/03036758.2019.1609052.

[70] E.E.R.I.EERI , “ Earthquake engineering research institute,” Web Page. Accessed: 2025. [Online]. Available: https://www.eeri.org/.

[71] E.E.R.I.EERI , “ M 7.1 - 1 km S of Matzaco, Mexico,” Web Page. Accessed: 2025. [Online]. Available: https://earthquake.usgs.gov/earthquakes/eventpage/us2000ar20/map.

[72] E.E.R.I.EERI , “ Ferndale, California-Learning from earthquakes.” Accessed: 2025. [Online]. Available: https://learningfromearthquakes.org/earthquakes/2022-12-20-ferndale-ca/?_observation_category=buildings&_paged=3#home.

[73] U.S. Geological Survey , “ Earthquake lists, maps, and statistics.” Accessed: 2025. [Online]. Available: https://www.usgs.gov/natural-hazards/earthquake-hazards/lists-maps-and-statistics.

[74] Pineda-Porras OA, Ordaz M. Seismic damage estimation in buried pipelines due to future earthquakes-The case of the Mexico City water system. Earthquake-resistant structures-Design, assessment and rehabilitation. IntechOpen; 2012. p. 131-50.

[75] Jaimes MA, Suárez G. Estimation of damage scenarios in the Mexico City basin caused by local crustal earthquakes. Bull Seismol Soc Am 2025; 115(4): 1907-23. https://doi.org/10.1785/0120240167.

[76] Velasquez HA, Hurtado Á, Soto D, Angel C. Impact of structural degradation on the seismic resilience of self-built housing: an approach based on intelligent circular resilience. J Build Pathol Rehabil 2025; 10(2): 123. https://doi.org/10.1007/s41024-025-00633-5.