<p>Reinforced concrete-steel composite systems often rely on headed stud shear connectors to transfer forces in solid slab applications. These components play a significant role in structural integrity, especially in scenarios involving high loads. This study collects 224 experimental test results to develop statistical and data-driven methods that predict the shear capacity of headed studs. In this database, the mean shear capacities ranged from approximately 62 kN to nearly 319 kN. Within the study context, a total of 14 machine learning models, including traditional linear and advanced ensemble regressors, are investigated. The study results indicated that advanced ensemble models demonstrate better predictive accuracy than linear and regularized approaches. The best-performing models achieved R² values as high as 0.98 on the test subset and root mean square errors below 11 kN. The feature importance analyses highlighted the diameter of the stud shank and the diameter of the weld collar as primary factors influencing capacity predictions, while the tensile strength of studs showed lower relevance. The partial dependence plots confirmed nonlinear relationships between input variables and predicted capacities. The comprehensive framework and findings of this study are expected to guide and help scientists and engineers understand and predict stud shear performance.</p>