Machine learning in earthquake engineering: A review on recent progress and future trends in seismic performance evaluation and design
Shuling Hu, Tong Guo, M. Shahria Alam, Yuji Koetaka, Elyas Ghafoori, Theodoros L. Karavasilis
Abstract
Applying machine learning (ML) in earthquake engineering has introduced new opportunities for better predicting, evaluating, and mitigating structural damage under seismic hazards . The rapid advancement of ML in earthquake engineering necessitates a thorough understanding of its potential and limitations to guide future research and practical applications effectively. This literature review focuses on the recent advancements of ML in structural seismic performance evaluation and design optimization. This paper comprehensively explores recent trends and innovations for each area, highlights ongoing challenges, and suggests future directions involving emerging technologies. Key findings reveal significant progress in ML methodologies. Still, challenges related to the accurate prediction of nonlinear hysteretic responses, the need for improved generalizability of ML models, the scarcity of high-quality data, effective feature selection techniques, and regional scale investigations remain. Moreover, the future research needs and strategies for addressing these challenges are presented.