To estimate uniaxial compressive strength of some sedimentary rock types from a set of index test results, selected soft computing methods were employed. Consequent comparative performances of these methods were also evaluated. Some selected sedimentary rock types (i.e., grainstone, wackestone–mudstone, boundstone, gypsum and silty marl) collected from the Qom Formation, in the vicinity of Qom city, central Iran, were examined. Four indices (i.e., block punch index, point load strength, Schmidt hammer rebound number and ultrasonic P-wave velocity) and uniaxial compressive strength (UCS) of these samples were determined. A multidisciplinary approach including multiple linear regression analysis-, fuzzy inference system-, artificial neural network- and adaptive neurofuzzy inference system-based models were used to estimate UCS from these index test results. Various statistical parameters (VAF, RMSE and R2) were determined to check the predictive performances of these models. The RMSE, VAF and R2 values for ANFIS model were 0.99, 3.77 and 0.99, respectively. In a comparative sense, ANFIS gives better predictive results. Accordingly, this method is being inferred as a predictive approach to estimate preliminary rock engineering purposes in comparison with alternate methods.
