Simulation and modeling study demonstrates machine-learning prediction of earthquake building damage across urban districts, highlighting utility for seismic emergency planning.
Rapid and reliable prediction of earthquake-induced building damage is crucial for assessing seismic risk and emergency management at the district level. This study developes a data-driven, scenario-based framework integrating regional hazard characterization, exposure modeling, physics-based nonlinear simulations, machine-learning damage prediction, economic loss estimation, and district-level risk profiling. Applied to Tehran under North Tehran Fault and South Ray Fault scenarios, a Random Forest model was trained on 409,590 nonlinear simulation samples to predict ATC-20 damage states. Internal and external validation achieved accuracies of 70.8% and 64.9%, respectively, providing an interpretable framework for district-level seismic risk prioritization and emergency planning.
No takes yet. Share an insight, caveat, or question.
Monirvaghefi et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: