Detecting radiometric outliers in airborne gamma-ray spectrometry data with isolation forest: A case study of REE-bearing rocks in central-north Brazil
Adolfo Barbosa da Silva, Geological Survey of Brazil, Rafael Pires de Lima · 2024
We show that Isolation Forest (IF) is an effective tool for radiometric outlier identification. By applying IF on airborne gamma-ray spectrometry data recorded in a region where Rare Earth Elements (REE) rich rocks have been mapped in central-north portion of Brazil, we found that IF scores is able to identify K, eTh, and eU outliers simultaneously and to map new targets for REE prospecting or associated mineralization.