Estimation of the undrained shear strength of sensitive clays using optimized inference intelligence system

The undrained shear strength of the sensitive clays is an important parameter for the design of the foundation of the civil engineering structures. In this study, novel hybrid machine learning approaches, namely ANFIS-CA and ANFIS-PSO, are developed to predict the undrained shear strength of the sensitive clays. These approaches are based on adaptive neuro-fuzzy inference system (ANFIS) and two metaheuristic optimizations techniques including cultural algorithm (CA) and particle swarm optimization (PSO). Unlike other empirical methods that relied on accurate determination of the pre-consolidation pressure, the proposed approaches are based on five reliable input parameters: depth, effective vertical stress, natural water content, liquid limit, and plastic limit. For this purpose, data of 216 sensitive clay samples obtained from different parts of Southern Finland were used for validating and training models. Standard statistical measures were used to evaluate performance of the models. The results show that the proposed hybrid ANFIS-PSO model obtained reasonably good accuracy (correlation coefficient: R = 0.715), in comparison with ANFIS-CA model (R = 0.6) in predicting the undrained shear strength of the sensitive clays. Therefore, the ANFIS-PSO model is very promising to predict the undrained shear strength of the sensitive clays with limited input parameters.