Abstract
This paper presents a comparative analysis of the Ordinary Kriging and Radial Basis Function (RBF) methods for interpolating gold (Au) concentrations in the territory of East Kazakhstan. The study is based on geochemical data comprising more than 800 observation points. To assess prediction accuracy, a portion of the original dataset was intentionally excluded and used as a control sample. Interpolation was performed on a unified regular computational grid containing approximately 50,000 nodes, which ensured the comparability of the results obtained by both methods. The quality of interpolation was evaluated using standard error metrics, namely Mean Absolute Error (MAE) and Root Mean Square Error (RMSE). The results show that the RBF method provides higher prediction accuracy compared to Ordinary Kriging. For the RBF method, MAE and RMSE values of 0.044324 ppm and 0.057145 ppm were obtained, respectively, whereas for the Kriging method the MAE was 0.071336 ppm and the RMSE was 0.082134 ppm. The higher efficiency of the RBF method is explained by its ability to flexibly approximate local variations of the geochemical field without strict assumptions about data stationarity. The obtained results confirm the suitability of the RBF method for interpolating geochemical data and identifying prospective anomalous zones under conditions of sparse and heterogeneous observation networks.