@article{scholars13825,
            note = {cited By 1},
            year = {2020},
          number = {2},
         journal = {Petroleum and Coal},
           pages = {525--541},
           title = {Litho-fluid facies modeling by using logistic regression},
       publisher = {Slovnaft VURUP a.s},
          volume = {62},
             url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-85087302087&partnerID=40&md5=ea2104111b84c613e410d1b7f199081c},
            issn = {13377027},
        abstract = {Rock-physics monitoring plays a crucial role in forecasting the facies distribution within the subsurface. However, litho-fluid estimation is challenging due to the randomness of the relationship between reservoir properties and facies distribution. As a result, some facies show similar responses to rock properties and attributes. Data-quality, as well as the elasticity difference between different facies, are the main factors affecting the efficiency of facies-discrimination. This study aims at predicting the distribution of gas-sand, wet-sand, and shale from elastic properties by using logistic regression. The effect of sand and gas distributions on fourteen elastic attributes has been tested to reduce the facies model's variables and determine the best lithology and fluid predictors. The results show that the near and far elastic impedances are the best lithology predictors, while the best fluid predictors are the Mu-Rho, Lambda-Rho/Mu-Rho, and Poisson's ratio. Accordingly, the coefficients of both models have been estimated by Markov-Chain Monte-Carlo simulation to calculate the sand and gas probabilities. Eventually, the two models have been merged by using appropriate cut-offs to turn the probabilities into facies estimates. The correct-classification-rate values of the gas-sand, wet-sand, and shale are 0.92, 0.81, and 0.91, respectively. {\^A}{\copyright} 2020 Slovnaft VURUP a.s.},
          author = {Gouda, M. F. and Salim, A. M. A.}
}