a. davatgar

Since most of the geochemical data analysis procedures require preliminary assumptions that lead to constraint and errors in the true nature of the data causing reduction in the effectiveness of the adopted methods. Therefore, a model-based clustering method as self-organizing map (SOM) were employed with the aim of recognizing Cu-Au mineralized zones by establishing an optimized exploration tool in Valezir area, NW of Iran. SOM as a dimension reduction method was introduced to recognize geochemical distribution patterns of Cu-Au with higher certainty while preserving the originality of the data. Subsequent to data preprocessing and testing different SOM architectures, an appropriate structure with a pattern containing six clusters was selected. Accordingly, the related elements distribution model was extracted and interpretation of the geochemical system represents two significant sets of elements in clusters (i.e. 1st, 2nd and 6th clusters) to anticipate the mechanism of distribution: 1- Copper and pertaining trace elements formation from intermediate to acidic hydrothermal solutions, which are localized in the northern part of the area and emplaced in the quartz monzo-dioritd intrusive body. 2- Au Anomalies and its associated elements As, Hg and Bi depicted in 2nd cluster. The Au anomalies follow geochemical pattern with Bi, Sb, As, and W that are mostly elongated from NW to SW of the area. It seems relatively the low enrichment of gold has occurred during the intrusion of the igneous body into older volcanic units that caused extensive alterations, remobilization and localization of Au and related elements. To assess the SOM results, a comparative study was carried out with the results obtained from hierarchical clustering analysis (HCA). The results illustrated higher performance by SOM approach in characterizing geochemical system and detecting the elements paragenetic sequence in the area for locating the exploration targets.