University of Tehran Civil Engineering Infrastructures Journal 2322-2093 59 1 2026 06 01 Integrating Machine Learning and Genetic Expression Programming for Enhanced Punching Shear Strength Prediction 139 160 99317 10.22059/ceij.2024.380171.2118 EN Alireza Mahmoudian M.Sc., Department of Civil Engineering, Shahid Rajaee Teacher Training University, Tehran, Iran. 0009-0005-4870-5493 Nima Tajik Ph.D. Candidate, Department of Civil, Structural and Environmental Engineering, State University of New York at Buffalo, USA. 0009-0006-5228-9223 Amirhossein Darabi M.Sc., School of Civil Engineering, Iran University of Science and Technology, Tehran, Iran. 0009-0000-2994-0205 Mostafa Mohammadzadeh Taleshi Ph.D. Candidate, Civil and Environmental Engineering Department, University of Nevada, Reno. 0009-0008-7049-5068 Saba Marmarchinia Ph.D. Candidate, Department of Civil, Structural and Environmental Engineering, State University of New York at Buffalo, USA. 0009-0000-4221-5200 Abazar Asghari Associate Professor, School of Civil Engineering, College of Engineering, University of Tehran, Iran. 0009-0005-3224-1815 Seyed Rasoul Mirghaderi Professor, School of Civil Engineering, College of Engineering, University of Tehran, Iran. 0009-0001-6256-3971 Journal Article 2024 07 30 Estimating the punching shear strength of Reinforced Concrete (RC) flat slabs is critical in structural engineering due to potential catastrophic failures. This study introduces advanced data-driven methods, including Machine Learning (ML), Deep Learning (DL), and Genetic Expression Programming (GEP), to improve predictions of punching shear strength. Analyzing a dataset of 380 test samples, the research evaluates various models such as linear regression, stochastic gradient descent, ridge regression, decision trees, K-nearest neighbors, random forests, adaptive boosting, Extreme Gradient Boosting (XGBoost) for ML, alongside Artificial Neural Networks (ANNs) for DL, and GEP for deriving explicit equations. Significant enhancements in model performance were achieved through rigorous hyperparameter tuning, notably with the XGBoost model, which attained an coefficient of determination (R²) score of 0.98, surpassing other models and existing code-based predictions. The study uses SHapley values to interpret model predictions, highlighting the significant impact of slab depth on punching shear strength, especially in the XGBoost model. Additionally, the GEP method derives explicit equations that accurately represent the relationship between input features and punching shear strength. This research highlights the advantages of advanced computational models and offers new insights into the factors influencing punching shear strength in RC slabs. https://ceij.ut.ac.ir/article_99317_cc5c572f4f6b9a63939c02b68e7f4cd9.pdf