基于斜方辉石成分的机器学习模型判别基性-超基性岩成因类型与成矿潜力

Key Points

• The research aims to utilize pyroxene compositions to classify the origin and mineral potential of basic-ultrabasic rocks.
• Compiled chemical data of pyroxene from global basic-ultrabasic rock bodies and mineral-rich rocks.
• Utilized classification algorithms to analyze pyroxene characteristics across different rock types.
• Compared features of pyroxene in relation to nickel-copper deposits and platinum group elements.
• Established that pyroxene characteristics differ among rock types using binary diagrams.
• Found high accuracy in origin classification using a combined model based on key elements like Ca, Cr, and Mg.
• Achieved over 90% accuracy, F1 score, and recall rate for distinguishing mineral potential using random forest and voting ensemble models.