Document Type : Original Research Paper
Authors
1 Department of Mining Engineering, Arak University of Technology, Arak, Iran.
2 Department of Mining in Arak university of Technology
Abstract
A total of 400 stream sediment samples were analyzed for 13 elements, and stepwise factor analysis was employed to generate geochemical maps indicative of mineralization. This method was utilized to develop a Geochemical Mineralization Probabilistic Index (GMPI) through a novel approach that produces geochemical evidence maps derived from stream sediment data. The study comprised a three-stage factor analysis of geochemical data collected from the Khomain Dehno region. The first factor included Zn, Pb, As, and Cd, accounting for 41.63% of the variance. The second factor comprised Mn, Mo, and Zr, explaining 21.86% of the variance, while the third factor consisted of Fe, Cu, and Ti, representing 7.79% of the variance. The cumulative variance explained by these three factors was 81%. Furthermore, a novel intelligent methodology, termed Relevant Vector Regression (RVR), enhanced with Cocoa Search (CS) and Harmony Search (HS) algorithms, is proposed for the prediction of the GMPI. The HS and CS algorithms were integrated with the RVR model to optimize its hyperparameters. In these models, Zn, Pb, As, and Cd served as input variables, while the GMPI was designated as the output variable. The performance of the predictive models was evaluated using Mean Squared Error (MSE) and the Coefficient of Determination (R²). The results indicated that the RVR model optimized with the HS algorithm exhibits superior performance, achieving an R² value of 0.99256 and an MSE of 0.0031455. These findings underscore the efficacy of the proposed approach for accurate GMPI estimation.
Keywords
- Stream sediments
- Stepwise factor analysis
- Relevant vector regression
- Cocoa search and Harmony search algorithms
Main Subjects
)