LAUSR

Abstract This research provides an innovative combination of an adaptive neuro-fuzzy inference system (ANFIS) model for solving a nonlinear and complex problem related to soil shear strength prediction. The new hybrid model is optimized by an optimization technique i.e., Henry gas solubility optimization (HGSO), called as HGSO-ANFIS. In predicting soil shear strength, the results of liquid limit, specific gravity, clay content, moisture content, void ratio, and plastic limit were considered and used as the model predictors. The HGSO-ANFIS model is implemented based on Henry’s law and can be used in engineering issues. The HGSO algorithm is developed based on the huddling behavior of gas to find the main answers and to avoid being trapped in the local minima. The search space in this model can be presented with a better performance than the base model. The performance of the new hybrid HGSO-ANFIS model was tested with real data to compare the other ANFIS-based models. The performance of the best HGSO-ANFIS model for the testing data was 0.954 and 0.1891 for coefficient of determination (R2) and root mean square error (RMSE), respectively. The model results showed that the new hybrid HGSO-ANFIS model can get higher level of accuracy compared to the other ANFIS-based models and it can be applied for various prediction and optimization problems.