Geometallurgical sampling has an important role in the geometallurgical programme. There are some critical aspects to be considered in geometallurgical sampling. Cluster analysis (CA) is one of the most popular methods that aids in creating domains. This study aims to use CA to create Geological domaining at the Sungun porphyry copper deposit (SPCD). Geological domains of SPCD were defined using PCA and K-means clustering with one hot encoding algorithm. Categorical data were encoded with one hot method and then PCA was used to condense the large new data set to its relevant features. In this research, we have exploited two validity indices to define the optimum number of clusters, namely the silhouette index and the elbow method. Therefore, clustering was performed using four clusters and geological domains were partitioned. The richest domain in this deposit is the second cluster in which the average grades of Cu and Mo are higher than in other clusters.
Keywords: Cluster analysis, geological domains, K-means, PCA, Sungun deposit;