Comparison of Clustering Methods in Mineral Potential Mapping of Barika Anomaly, West Azerbaijan

Nowadays GIS techniques are used as a conventional tool for integrating geographic information datasets. In these methods, integration is done according to quality and quantity of datasets and using appropriate weighting approaches. Finally, with Classification methods like Geometric Distances, Mineral Potential Maps (MPM) is produced. By increasing application in data processing, clustering methods classify samples into groups by similarity between them. In this paper, K_ Means and Fuzzy K_ Means clustering methods are discussed for mapping potential zones of Gold mineralization, then the results are compared with GIS method, Index Overlaying, for the Barika area in 1:100,000 Alut Sheet in South of Azerbaijan. In the Barika area, information of drilling points aren’t available, so it’s not possible to determine number of classes and boundaries of each class for final score map, but in clustering methods, the optimum number of class for output map is done automatically and is tested for the Barika Anomaly data. The results show that clustering methods need no weighting and it’s possible to use it with low information than GIS_ based method. The results also show that both of approaches, clustering methods and Index overlaying, display almost an equal area for the most potential zone, however clustering methods need low information for analyzing, but in Index overlaying, it is needed to have more information for weighting and determining threshold for classification of final scores.