improvement of power spectrum-area fractal model using linear discriminant analysis in frequency domain of geochemical data

the methods and techniques for separation of geochemical anomaly, background and noise have been the issue of research for many years. several people have applied the power spectrum-area (sa) fractal model to identify geochemical anomalies, background and noises. the purpose of s-a method is to divide the power spectrum (ps) values into components characterized by similar scaling properties. the high frequencies in surface geochemical distribution map are related to surface mineral deposits and geochemical noises. very low frequencies are related to background values and very deep mineral deposits. s-a fractal method can be used to identify the deep and blind mineral deposits. very low frequencies in surface geochemical data that have direct relationship to high pss can be separated using s-a fractal method. in this study, investigations have been conducted on dalli cu-au porphyry deposit that is located in the central province of iran. three ranges and two cutoff values for ps have been identified for cu using s-a model. this paper introduces the combined approach of s-a fractal method and linear discriminant analysis (lda) in frequency domain of geochemical data to improve hese threshold values. the lda is a supervised classification method, which investigates a relationship between features space and a dependent categorical class using a set of training data and be able to verify the results of cluster analysis. in this study, lda method has been performed on the results of the s-a method then the confusion matrix has been formed and correct classification rate indexes (ccr) have been calculated in different threshold values. then the results were compared together and the optimum threshold value between low and moderate ps values were determined. the results of s-a fractal method have been properly improved using the proposed combined method of s-a fractal and lda.