Utilizing advanced machine learning approaches to assess the seismic fragility of non-engineered masonry structures

• DOI: 10.1016/j.rineng.2024.101750
• Publication Date: 2024-01-05T09:08:36Z
• Authors: Ehsan Harirchian; Seyed Ehsan Aghak...; Viviana Novelli; Tom Lahmer; Shahla Rasulzade
• Keywords: Resilience

Seismic fragility assessment provides a substantial tool for assessing the seismic resilience of these buildings. However, using traditional numerical methods to derive curves poses significant challenges. These often overlook diverse range buildings found in different regions, as they rely on standardized assumptions and parameters. Consequently, may not accurately capture response various building types. Alternatively, extensive data collection becomes essential address this knowledge gap by understanding local construction techniques identifying relevant This is crucial developing reliable analytical approaches that can curves. To overcome challenges, research employs four Machine Learning (ML) techniques, namely Support Vector Regression (SVR), Stochastic Gradient Descent (SGD), Random Forest (RF), Linear (LR), probability collapse terms Peak Ground Acceleration (PGA). achieve objective, comprehensive input/output dataset consisting on-site collected from 646 masonry walls Malawi used. Adopted ML models are trained tested entire then again only most highly correlated features. The study includes comparative analysis efficiency accuracy each approach influence used analyses. (RF) technique emerges efficient deriving surveyed achieved lowest values evaluation metrics methods. scored Mean Absolute Percentage Error (MAPE) 16.8 %, Root Square (RMSE) 0.0547. results highlight potential particularly RF, derivation with proper levels accuracy.