Dataset of surrogate modeling and optimization for bolted timber joints in glulam structures

This dataset supports the findings of the study titled “Data-driven and interpretable surrogate modeling for mechanical performance prediction of bolted timber joints in long-span glulam structures”.The dataset includes experimental data, finite element simulation results, and surrogate modeling data used to investigate the semi-rigid mechanical behavior of bolted timber joints.Specifically, the dataset contains:(1) Full-scale experimental test data of bolted timber joints, including moment–rotation (M–θ) curves;(2) Validated finite element simulation results calibrated against experimental observations;(3) Input–output datasets used for training and testing surrogate models (SVR, Random Forest, XGBoost, and MLP-BP);(4) Prediction results of surrogate models for key mechanical performance indicators, including strength, initial stiffness, and ductility;(5) Multi-objective optimization results obtained using NSGA-II, including Pareto-optimal solution sets.The surrogate models are further interpreted using SHAP (SHapley Additive exPlanations) to quantify the contribution of input parameters to the predicted responses.All data are provided in accessible formats (e.g., CSV, MAT, or XLSX) and can be used to reproduce the results presented in the associated publication.This dataset is intended for researchers and engineers working on timber structures, structural connections, and data-driven structural design.