Machine Learning model interpretability using SHAP values: Application to Igneous Rock Classification task

• Published in: Applied computing and geosciences Vol. 23; p. 100178
• Main Authors: Antonini, Antonella S., Tanzola, Juan, Asiain, Lucía, Ferracutti, Gabriela R., Castro, Silvia M., Bjerg, Ernesto A., Ganuza, María Luján
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01-09-2024; Elsevier
• Subjects: Explainable artificial intelligence; Geochemical discrimination; Mafic–ultramafic rocks; Random forest classification; SHAP values; Explainable artificial intelligence; Mafic–ultramafic rocks; Random forest classification; Geochemical discrimination; SHAP values
• ISSN: 2590-1974; 2590-1974
• DOI: 10.1016/j.acags.2024.100178

El Fierro intrusive body is one of the bodies that compose the La Jovita–Las Aguilas mafic–ultramafic belt, located in the Sierra Grande de San Luis, Argentina. The units of this belt carry a base metal sulfide (BMS) mineralization and platinum group minerals (PGM). The macroscopic description of mafic and ultramafic rocks, as is usually done by the mining exploration companies, leads to an imprecise modal classification of the rocks. In this study, we develop a random forest-based prediction model, which uses geochemical parameters to classify mafic and ultramafic rocks intercepted by drill cores. This model showed an accuracy of between 86% and 94%, and an f1_score of 96%. Random forest classification is a widely adopted Machine Learning approach to construct predictive models across various research domains. However, as models become more complex, their interpretation can be considerably difficult. To interpret the model results, we use both global and local perspectives, incorporating the SHAP (SHapley Additive exPlanations) method. The SHAP technique allows us to analyze individual samples using force plots, and provides a measure of the importance of each geochemical input attribute in the model output. As a result of analyzing the contribution of each input feature to the model, the three variables with the highest contributions were identified in the following order: Al2O3, MgO, and Sr.
•The study aims to classify mafic and ultramafic rocks using a Random Forest model with geochemical parameters.•While the Random Forest model is widely used for predictive modeling, interpreting complex models can be challenging.•The SHAP technique measures the importance of input geochemical attributes for interpreting model results globally and locally.