Uniaxial compressive strength (UCS) and modulus of elasticity (Es) are two basic parameters in geotechnical engineering. Determination of UCS and Es requires considerable time and money to collect and prepare suitable specimens and laboratory tests. The aim of this study was to evaluate the methods of multilayer perceptron artificial neural network (MLP-ANN), support vector regression (SVR), simple regression (SR) and multivariate regression analysis (MRA) to estimate UCS and Es of the carbonate samples at Bakhtiari dam site, southwest of Iran. The P-wave velocity (Vp), porosity, density, and water absorption percentage were used as input parameters in methods. The results of petrographic studies showed that by changing the texture of the samples from Wackestone to Grainstone, the UCS and Es of the samples increase and the porosity decreases. The MRA method predicted UCS and Es with determination coefficients of 0.72 and 0.75, respectively. The use of MLP-ANN improved the determination coefficients to level of R2 = 0.83 for UCS and R2 = 0.92 for Es. The SVR forecasted the UCS and Es with coefficients of determination of 0.89 and 0.97, respectively. Various criteria were used to evaluate the performance of MLP-ANN, SR and MRA methods. Results showed that the SVR method performed better than the SR, MLP-ANN and MRA methods in predicting the UCS and Es. Previous researchers’ relationships were evaluated and results showed that most of previous relationships could be used to estimate the static properties.
