An Investigation of Pb Geochemical Behavior Respect to Those of Fe and Zn Based on k-Means Clustering Method


A well-known algorithm of clustering is k-means by which the data are divided into k classes based upon a distance criterion. In the present research, by using k-means method for classifying data derived from exploration boreholes in the Parkam deposit, the optimum k has been calculated and then the data have been clustered and the relative geochemical behavioral characteristics analyzed. The criterion used for determining the optimum k ranged the number of classes from k=3 to k=10 and afterwards, analyzed derived classifications in order to choose the optimum k. Results showed that clustering with k=3 in case of Pb and Zn and k=4 in case of Pb and Fe were better than other classes number in each case and according to derived classification and above cases, increase in Pb grade is followed by increase in Zn grade and also There is an increase and subsequent decrease in Fe grade. Thus, it is possible to investigate the fluctuation of elements such as Cu or Pb with other elements existing in done analysis using suggested above method that can provide a very appropriate viewpoint in front of this industry decision makers.