loading page

Classification for the Determination of Estimation Domains in a Cu-Zn Skarn Deposit in Central Peru, New Approach using Gaussian Kernel Support Vector Machine
  • Harold Velasquez Sanchez,
  • Marjory Aguilar
Harold Velasquez Sanchez
Universidad Nacional de Ingeniería

Corresponding Author:[email protected]

Author Profile
Marjory Aguilar
Universidad Católica de Santa María
Author Profile


The Ore-control block-model in an open pit mine constitutes the final outcome after rigorous analysis and interpretations of a wide range of geological data. Moreover, modeling different variables in current tools such as GIS software or specialized programs may be highly time consuming and these software’s and tools may restrict you to use determined types of data and can be un-accurate when they are utilized for making attempts to find out multivariate relationships. One of these ore-control tasks that has to be done is the determination of the short-term grades for the block-model, within which diverse mathematical calculations are carried out. In order to get the grade estimation, geologists require to determine the Estimation Domain for every lithology and then go forward with the estimation techniques using the laboratory grades from blastholes samples as an input, so it is clear that estimation domains are essential for ore-control purposes. Estimation domains require logged lithology and grades input of every blast hole sample; the logged lithology is directly obtained by the geology staff by describing the detritus from blastholes. This paper aims to present novel results in determining the relationships of multivariate laboratory assays in a Cu-Zn Skarn deposit and its corresponding logged lithology using kernel support vector machine algorithm, so with this, it may be possible for geologists to forecast lithology of every sample, based on its chemical content, and so, they will be able to determine estimation domains.