Enhancing of uniaxial compressive strength of travertine rock prediction through machine learning and multivariate analysis

• DOI: 10.1016/j.rineng.2023.101593
• Publication Date: 2023-11-18T01:07:12Z
• Authors: Dima A. Malkawi; Samer R. Rabab'ah; Abdulla A. Sharo; Hussein Aldeeky; Ghada K. Al-Souliman; Haitham O. Saleh
• Keywords: Schmidt hammer

Indirect methods for predicting material properties in rock engineering are vital assessing elastic mechanical properties. Accurately holds significant importance and geotechnical engineering, as it strongly influences decisions about the design construction of infrastructure projects. Uniaxial compressive strength (UCS) is one most important understanding how rocks geological formations respond to stress deformation. However, standard UCS test faces several challenges, including its destructive nature, high costs, time-consuming procedures, requirement high-quality samples. Therefore, there a growing demand indirect estimate UCS, which invaluable tools evaluating materials. The study aimed comprehensively analyze relationships between travertine samples collected from Dead Sea Jordan Valley seven different indices by utilizing parametric non-parametric methods. laboratory results indicate that area's possesses desirable reveal certain indices, such Schmidt hammer, Leeb rebound hardness, Point Load, correlate with Compressive Strength (UCS). Conversely, other specifically dry density, absorption, pulse velocity, porosity, exhibit considerably weaker or very weak relationship UCS. paper employs three machine learning techniques, namely Tree model, k-nearest neighbors (KNN), Artificial Neural Networks (ANN), develop predictive models strength. were trained on dataset corresponding values. study's revealed M5 tree model suitable method It demonstrates robust performance across spectrum metrics boasts low prediction errors. Following KNN, ANN, regression descending order performance.