RC shear walls are commonly used as lateral load-resisting elements in seismic regions, and the estimation of their shear strengths can become simultaneously design-critical and complex when they have so-called squat geometries, i.e., height-to-length ratios less than two. This paper presents a study on the training and interpretation of an advanced machine-learning model that strategically combines two algorithms for the said purpose. To train the model, a comprehensive shear strength database of 434 samples of squat RC walls is utilized. First, the eXtreme Gradient Boosting (XGBoost) algorithm is used to establish a predictive model for estimating the shear strength, wherein 70% and 30% of the data are respectively used for training and validation. This effort resulted in an approximately 97% validation accuracy, which well exceeds current mechanics-based/semiempirical models. Second, the SHapley Additive exPlanations (SHAP) algorithm is used to estimate the relative importance of the factors affecting XGBoost’s shear strength estimates. This step thus enabled physical and quantitative interpretations of the input-output dependencies, which are nominally hidden in conventional machine-learning approaches. Through this setup, several squat wall attributes are identified as being critical in shear strength estimates.

Inelastic response of reinforced concrete columns to combined axial and flexural loading is characterized by plastic deformations localized in small regions, which are idealized as plastic hinges. Under extreme events such as earthquakes, the load-carrying and deformation capacities of reinforced concrete beam/columns are highly dependent on the accuracy of this idealization for which the plastic hinge length is a key parameter. From a design perspective, a reinforced concrete column can only attain the ductility characteristics prescribed by its performance level if it is provided with sufficient confinement along the length of its plastic hinge zones. From an analysis standpoint, an efficient, nonlocalized, and objective finite-element simulation of column behavior requires accurate plastic hinge length definitions. This paper presents a novel data-driven model for predicting the plastic hinge length of reinforced concrete columns and its implementation in force-based fiber beam-column elements. The model is based on an ensemble machine learning algorithm named adaptive boosting (AdaBoost) and is trained using the results of 133 reinforced concrete column tests conducted in the period from 1984 to 2013. The performance of the model is assessed using the 10-fold cross-validation technique. It is shown that the prediction accuracy achieved using the proposed method is considerably higher than those of state-of-the-art empirical relationships and several other highly effective machine learning base models. Furthermore, numerical experiments reveal that the force-based beam-column models using plastic hinge length predictions of the developed model closely resemble the monotonic and cyclic behavior observed in laboratory experiments.