Comparison of Six Machine-Learning Methods for Predicting the Tensile Strength (Brazilian) of Evaporitic Rocks

Rock tensile strength (TS) is an important parameter for the initial design of engineering applications. The Brazilian (BTS) test suggested by International Society Mechanics and American Testing Materials widely used to assess TS rocks indirectly. Evaporitic rock blocks were collected from Al Ain city in United Arab Emirates. Samples tested, a database 48 samples was created. Although previous studies have applied different methods such as adaptive neuro-fuzzy inference system linear regression BTS prediction, we are not aware any study that employed regularization techniques, Elastic Net, Ridge, Lasso, Keras based sequential neural network models. These techniques powerful feature selection tools can prevent overfitting improve model performance prediction accuracy. In this study, six algorithms, namely, classical best subsets, three artificial networks with two application-programming interfaces (Keras on TensorFlow Neural Net) determine predictive BTS. models compared through ten-fold cross-validation. obtained results revealed outperformed all other considered