Engineering strength of fiber-reinforced soil estimated by swarm intelligence optimized regression system

Fiber-reinforced soil (FRS) has been used as a promising alternative material for civil and construction engineering. Shear strength of FRS is influenced complexly by many factors including fiber properties, soil properties, and stress conditions. This inherent complexity limits the ability of designers to assess shear strength parameters and has made it difficult to establish a mathematical model for accurately predicting the FRS shear strength. Accurately estimating the shear strength of FRS is vital for civil engineers in designing geotechnical structures and management. Thus, this work proposed a novel computational method, namely a swarm intelligence optimized regression (SIOR) system to estimate the peak shear strength of randomly distributed FRS. The SIOR system integrates a machine learning technique with an enhanced swarm intelligence algorithm to obtain reliable and good performance in prediction process. The real-world FRS dataset collected over the past 30 years was used to validate the proposed system. To demonstrate the capability of the proposed system, the SIOR modeling results were compared with those obtained using numeric predictive models. The analytical results confirm that the SIOR system is superior to a baseline machine learning model and empirical methods via cross-fold validation and hypothesis test with accuracy improvement from 44.7 to 99.7% in mean absolute percentage error. Therefore, the SIOR system can significantly improve the prediction accuracy and facilitate civil engineers in estimating the shear strength of the FRS.