Advanced predictive modeling for masonry walls: A comparative study of six AI models and existing empirical formulas

• This study presents innovative approaches to estimate the compressive strength of masonry wall systems. • Six novel AI approaches (ANN, GEP, RF, ETR, XGBoost, and AdaBoost) are employed. • AI approaches were compared with the predictions of existing seven different studies. • Key influencing factors affecting masonry compressive strength are identified Masonry walls are still valued for their affordability and endurance. However, analytical modeling for evaluating strength capacity is challenging due to their complex nonlinear behavior driven by their heterogeneous composition. The integration of Artificial Intelligence (AI) with advanced modeling techniques makes it possible to evaluate historical masonry structures more precisely. To address this, the study introduces innovative AI technologies aimed at redefining how masonry strength is estimated. Six novel AI approaches, namely Artificial Neural Network (ANN), Gene Expression Programming (GEP), Random Forest (RF), Extra Trees Regressor (ETR), Extreme Gradient Boosting (XGBoost) and AdaBoost regression models, were employed. By leveraging a comprehensive dataset of 452 full-scale masonry specimens, the study evaluates the influential parameters governing compressive strength, unveiling intricate relationships that traditional models often overlook. Furthermore, the superiority of the developed models is demonstrated through a rigorous comparative analysis against seven existing predictive expressions from the literature, revealing a remarkable improvement in accuracy and reliability. Among the AI models, ensemble learning techniques exhibit exceptional generalization performance. Based on the results, RF and ETR are the most accurate models, with exceptional performance metrics R² = 0.968 and 0.967, respectively, with minimal errors. This study shows that AI-based modeling can effectively address the limitations of experimental methods, thereby enhancing the applicability of results in engineering contexts and contributing to advancements in masonry research.