成都理工大学学报（自然科学版）2025，Vol.52Issue(5)：844-858,15.DOI:10.12474/cdlgzrkx.2025071901
机器学习在矿物岩石地球化学大数据挖掘中的应用与展望
Applications and perspectives of machine learning in geochemical big data mining of minerals and rocks
摘要
Abstract
With the rapid advancement of information technology,geoscience research has entered the data-intensive era,where the large-scale mining of mineral and rock geochemical data has become a critical approach for deciphering geological evolution,deepening our understanding of ore-forming processes and enhancing exploration efficiency.As a representative data-driven modeling technique,machine learning offers powerful capabilities for identifying hidden patterns and key features within complex datasets,thereby providing new methodological avenues for geoscientific studies.This paper systematically reviews the general workflow of mineral-rock geochemical big data mining;outlines the basic principles of representative machine learning algorithms;and evaluates their recent applications in tectonic setting discrimination,rock genesis and evolutionary reconstruction,mineral prospectivity assessment,and ore deposit type classification.On this basis,we have summarized the strengths and limitations of machine learning in geochemical data mining,highlighting challenges such as limited sample sizes,uneven data distribution,and insufficient model interpretability.Furthermore,we discuss future research prospects in algorithm optimization,the development of mineral-rock geochemical databases,the integration of deep learning and transfer learning,multisource data fusion,interpretable artificial intelligence,and the application of low-code frameworks.The widespread application of machine learning in mineral and rock geochemical data mining is expected to provide robust theoretical support and methodological pathways for the quantitative analysis of geological processes,intelligent decision-making in mineral exploration,and the systematic development of Earth science research.关键词
机器学习/矿物岩石地球化学/大数据挖掘/成矿潜力评价/矿床成因判别Key words
machine learning/mineral and rock geochemistry/big data mining/mineral prospectivity assessment/ore deposit type classification分类
天文与地球科学引用本文复制引用
王智宇,王达,邱昆峰,张晓暄,才艺伟,王明阳..机器学习在矿物岩石地球化学大数据挖掘中的应用与展望[J].成都理工大学学报（自然科学版）,2025,52(5):844-858,15.基金项目
地球深部探测与矿产资源勘查国家科技重大专项(2024ZD1001607) （2024ZD1001607）
地质过程与成矿预测全国重点实验室开放课题项目(GPMR202522). （GPMR202522）
