Peak-Based Nuclide Identification in HPGe $\gamma$-Spectrometry with Machine Learning and SHAP
Researchers have developed a machine learning model to automate nuclide identification in high-purity germanium gamma spectra, a process typically requiring significant expert time. The model, trained on 65 isotopes, achieved an F1 score of 0.97, outperforming traditional software which scored 0.84. Shapley Additive Explanations were used to demonstrate that the model relies on physically relevant photopeaks for its predictions, validating its approach. AI