Prediction of reinforced concrete walls shear strength based on soft computing-based techniques

Abstract The precise estimation of the shear strength reinforced concrete walls is critical for structural engineers. This projection, nevertheless, exceedingly complicated because varied geometries, plethora load cases, and highly nonlinear relationships between design requirements strength. Recent related code regulations mostly depend on experimental formulations, which have a variety constraints establish low prediction accuracy. Hence, different soft computing techniques are used in this study to evaluate capacity walls. In particular, developed models estimating been investigated, based test data accessible relevant literature. Adaptive neuro-fuzzy inference system, integrated genetic algorithms, particle swarm optimization methods were optimize fuzzy model’s membership function range results compared outcomes random forests (RF) model. To determine accuracy models, assessed using several indices. Outliers anticipated identified replaced with appropriate values ensure comparison resulting findings demonstrates potential hybrid reliably effectively. revealed that RF model RMSE = 151.89, MAE 111.52, $${R}^2$$ R 2 0.9351 has best Integrated GAFIS PSOFIS performed virtually identically had fewer errors than ANFIS. sensitivity analysis shows thickness wall $$({b_\mathrm{{w}}})$$ ( b w ) compressive $$({f_\mathrm{{c}}})$$ f c most least effects strength, respectively.