ESTIMATION OF SOIL ELECTRICAL RESISTIVITY USING MLP, RBF, ANFIS AND SVM APPROACHES

The knowledge of soil electrical resistivity proves essential for a better earthing in order to ensure the protection of telecommunications and electrical energy networks. This study aims to estimate the value of the electrical resistivity of a site\'s soil from soil humidity and ambient temperature. The data used were measured at sites in the city of Lome and its surroundings. We developed models using Artificial Neural Network (precisely Multi-Layer Perceptron (MLP) and Radial Basis Function (RBF)), Adaptive Neuro-Fuzzy Inference System (ANFIS) and Support Vector Machine (SVM). The MAPE (Mean Absolute Percentage Error) errors obtained are 0.0011761% for the MLP model, 0.0719309% for the RBF model, 0.00105% for the ANFIS model and 2.89466% for the SVM model. We can say that the results are satisfactory for all models but the ANFIS model is better, given these performances compared to other models. The latter is then retained for the prediction of soil electrical resistivity.