A New Practical Formula for Bond-Slip Stress of Cold-Formed CFSTs Stub-Columns in a Post-fire Based on Four Cycles of Push-Out Tests Using Decision Tree-Based Machine Learning Techniques

In order to quantitatively evaluate the shear-bearing capacity of shear studs of square concrete-filled steel tubes (CFTs) at constant temperatures of 20 °C, 250 °C, 500 °C, and 750 °C, 16 CFTs square column specimens were constructed. To determine the task of the shear studs, three different types of them were designed and compared with the specimens without shear studs. Results indicated that the peak strain corresponding to the ultimate strength is increased by about 85% when applying shear studs. In the second part of this research, a formula is obtained from the model tree technique as the machine learning algorithms in order to predict the axial compression capacity CFTs specimens. Then, to assess the validity of the proposed model for predicting the axial compression capacity, artificial neural network model also was developed and the prediction results of two models were compared using various performance measures. The comparative study indicated that the proposed model tree achieved a superior prediction compared to artificial neural network in the prediction of the axial compression capacity of CFTs for practical engineering design.