Application of GMDH-Type neural network for path loss predictions

Abstract

This paper proposes the use of Group Method of Data Handling (GMDH) technique for Path Loss Prediction as a non-linear function approximation in  Cellular mobile network propagation losses. The paper compared prediction accuracy of GMDH with adaptive neuro-fuzzy inference systems (ANFIS) using four statistical performance indices with actual signal strength measurement taken at certain suburban areas of Bauchi metropolis, Nigeria. The  proposed GMDH model was found to offer improved prediction results in terms of reducing the root mean square error (RMSE) and mean absolute error  by 33.26% and 21.86% respectively over the ANFIS model.