We applied the Backward Elimination (BE) method as a criterion-based iterative stepwise method to estimate the concentration of Au from the variables Ag, Cu, Pb, and Zn. We optimized the quadratic polynomial model (QPM) on various boreholes and trenches. The results indicate that the vertical zonation of Au is associated with Ag and Cu, along with their respective elemental functions Ag×Cu, Pb2, Pb×Zn, and Zn2. On the other hand, the lateral dispersion of Au is determined by Pb and Zn, along with the essential functions Ag×Zn and Pb×Zn. Zonation is in-depth and is indicated primarily by Zn rather than Cu and minor Pb, with Cu-Pb-Zn gradually extending upward at the upper levels. The vertical zonation trend describes the Ag<Au<Cu<Pb<Zn sequence from proximal to distal veins, with the near-surface zone exhibiting enrichment in Zn. At the Glojeh site, the host rock value, Ag×Cu, Zn2, and Pb×Zn, were identified as strong indicators of the vertical occurrence of Au in the upper host rock, whereas Pb2 proved more suitable for the deeper boreholes. Cu→Pb→Zn demonstrates the enrichment trend from depth to the surface. The lateral kriging surface estimation confirms the dispersion pattern of elements revealed in-depth at Glojeh. Our aims with this approach were to reduce the cost of fire-assay analysis due to high precision estimation by associated variables and determine elemental interactions and the lateral or vertical dependencies of elements.