Using data-driven methods to select intensity measures for estimating the seismic response of cross-laminated timber walls

Abstract

Accurate identification of seismic intensity measures is important for the design and evaluation of the seismic response of structures. The research focused on selecting intensity measures for estimating the seismic response of cross‐laminated timber (CLT) walls using data-driven methods. Data-driven approaches were employed to analyze a large dataset from numerical structural analysis results. Various candidate intensity measures were selected using data-driven based feature selection processes and statistical regression models were applied to identify the most influential variables in estimating the nonlinear behavior of CLT wall systems under earthquakes. The findings offer valuable insights into the appropriate intensity measure in the evaluation and design of CLT-walled building systems, enhancing the effectiveness of performance-based structural evaluations under earthquakes.