Prediction effect of farmyard manure, multiple passes and moisture content on clay soil compaction using adaptive neuro-fuzzy inference system

Soil compaction by machine traffic is a complex process with many interacting factors. The strength of adaptive neuro-fuzzy inference system (ANFIS) is the ability to handle linguistic concepts and find nonlinear relationships between inputs and outputs parameters. In this research, the effect of farmyard manure, the number of tire passes, soil moisture contents and three average depths on clay soil compaction is predicted using ANFIS and Regression. For the prediction of soil compaction, an agricultural tractor tire was used and the experiments were carried out in the controlled condition of soil bin facility utilizing a well-equipped single-wheel tester. To measure soil compaction, cylindrical cores in groups of three were inserted into the three different depths. Various member function ANFIS were tested to discover the supervised ANFIS-based models for the soil compaction. On the basis of statistical performance criteria of MAPE and R2, Gaussian curve built-in membership function (gaussmf) was found as a proper model. In addition, the ANFIS model with 'Gaussian mf' is recommended considering the higher prediction performance values of MAPE = 0.2957%. The regression analyses of ANFI and Multiple Linear Regression (MLR) revealed a high correlation with farmyard manure, the number of tire passes, soil moisture, and depth. Also, it showed a higher performance compared to the regression model for predicting soil compaction. Thus, it can be concluded that ANFIS-based methodology is a soft computing approach that provides excellent nonlinear systems such as soil compaction.