MACHINE LEARNING-BASED PREDICTION OF SEISMICRESPONSE OF CONCRETE STRUCTURES WITH VISCOUSDAMPER

This paper presents a comprehensive study on the prediction of seismic responses, specifically theMaximum Roof Drift (MRD), of concrete structures equipped with viscous dampers using machinelearning (ML) models. Through a detailed feature importance analysis, key parameters influencingMRD were identified, highlighting the critical role of structural and seismic attributes. Six ML modelswere evaluated for their predictive accuracy, with ensemble methods like XGBoost and RandomForest demonstrating superior performance. The XGBoost model, in particular, showed exceptionalpredictive capabilities, as evidenced by its high R-squared values. This research underscores thepotential of ML in enhancing the prediction and understanding of seismic responses, offering valuableinsights for the design and retrofitting of resilient structures against seismic events.

Keywords:

Seismic Response , Viscous Damper , Machine Learning , Opensees