Prediction of Axial Compression Bearing Capacity of Built-In Steel Reinforced Concrete Filled Circular Steel Tube Based on Support Vector Machine

This paper developed a support vector machine (SVM) regression model to predict the axial compressive capacity of concrete-filled steel tubular (CFST) columns with built-in steel reinforcement. The input parameters of the regression model included the calculated length, outer diameter, wall thickness, and yield strength of the steel tube, the axial compressive strength of concrete, the cross-sectional area of the embedded steel reinforcement, and its yield strength. The output parameter was the experimentally measured axial compressive capacity. A dataset of 38 specimens was utilized, with 30 samples for model training and 8 for testing. The results demonstrated that the SVM regression model achieved a coefficient of determination (R²) of 0.98435, a mean absolute error (MAE) of 49.119, and a mean bias error (MBE) of -3.679 on the training set. For the testing set, the model yielded an R² of 0.9455, an MAE of 96.6133, and an MBE of -19.9404. These findings indicate that the proposed SVM model provides accurate predictions for the axial compressive capacity of CFST columns with built-in steel reinforcement, offering robust theoretical support and a reliable predictive tool for related engineering design and performance evaluation.