A neuro-fuzzy model for modulus of deformation of jointed rock masses

Use of indirect estimation methods for some rock mass parameter is considered traditionally in the rock mechanics applications. Generally, the regression based-statistical methods are used to develop an empirical equation. However, new techniques such as artificial neural networks, fuzzy inference systems or neuro-fuzzy systems were employed in recent years. In this study, construction of a neuro-fuzzy system to estimate the deformation modulus of rock masses is aimed, because this modulus has a crucial importance for many design approaches in rock engineering. For the purpose, a database including 115 data sets was employed and a neuro-fuzzy system consisting of two inputs, one output and three layers was constructed. After learning process, total 18 if-then fuzzy rules were obtained. The performance values such as RMSE, VAF, absolute error and coefficient of cross-correlation were calculated and, the constructed neuro-fuzzy model exhibited a high performance according to the performance indices. (C) 2004 Elsevier Ltd. All rights reserved.