Exploration and investments of deep ore deposits is demanded increasingly regarding to the limited explored surficial deposits. Among the exploration techniques, geochemical data processing is of high importance due to the related direct sampling and detecting the highly mobile elemental (geochemical variables) halos. Results of this processing techniques has been more reliable than before due to the development of analytical instruments and data processing techniques. Meanwhile, designing the drilling plan using surficial data is at high risk and might not be consistent with the deep ore body. Hence, optimal using of data and proper processing techniques decrease the operational risks and consequently expenses. In the preset thesis, training area and 75 percent accuracy in the experimental area. In the qualitative part, anomalies and the background are detected using discriminant Analysis, support vector machine (SVM) and random forests techniques. The mentioned techniques are categorized as supervised methods and their parameters are extendable to whole the study area using the specifications of a trained dataset of a known area. Results of SVM with Gaussian kernel is of the best accuracy value (87 percent) among the results of the mentioned methods. Considering the operative risks, SVM method is known as the proper and optimized one. After modeling in the two parts of qualitative and quantitative, results are extended to the whole area and further drilling targets are specified in North-Western parts of Zarshuran deposit.:
