Regimentation of geochemical indicator elements employing convolutional deep learning algorithm

A convolutional deep learning (CDL) algorithm has demonstrated high accuracy in clustering geochemical indicator elements, achieving root mean square error (RMSE) values below 20% and adjusted R² values above 90%. This approach significantly outperforms traditional methods for recognizing multi-element geochemical anomalies, particularly in mineral exploration contexts. Applied to the Takht-e Soleyman District in Iran, the algorithm effectively predicted dependent variables such as lead (Pb) and silver (Ag) using their independent variables, confirming its robustness and potential for broader geological studies.

Sabbaghi, H., & Tabatabaei, S. H. (2023). Regimentation of geochemical indicator elements employing convolutional deep learning algorithm. Frontiers in Environmental Science, 11. https://doi.org/10.3389/fenvs.2023.1076302