Mohammad Hossein Ghalehnoee - Faculty of Mining and Metallurgical Engineering, Yazd University, Yazd, Iran
Farhad Mohammad Torab - Faculty of Mining and Metallurgical Engineering, Yazd University, Yazd, Iran

A total 596 stream sediment samples have been carried out from Takab area which is one of the well-knownmineralization zones of northwest of Iran. Geochemical data are typical compositional data which should beopened prior to univariate and multivariate data analysis. In this study an ilr-transformation method has beenapplied to reduce the effect of outliers and censored values. Principle component analysis (PCA) has beenperformed based on ilr-transformed data and 6 significant PCs loaded. With the use of PC maps anomalous areasin the entire region have been suggested for more works such as geochemical rock sampling.

censored values, ilr-transformation, Ln-transformation, outlier, principle component analysis