摘要
花岗岩在全球广泛分布,不同成因类型的花岗岩记录了不同的构造环境、源区岩石类型或深部岩浆过程等,因此正确识别花岗岩成因类型(I型、S型、A型花岗岩)具有重要意义。磷灰石作为花岗岩中最常见的一种副矿物,蕴含多种主微量元素,记录着花岗岩的源区和岩浆物理化学特征,因此成为判别花岗岩成因类型的重要指标。本文收集了已发表的I型、S型和A型花岗岩的磷灰石微量元素数据,采用随机森林(Random Forest,RF)和支持向量机(Support Vector Machine,SVM)两种机器学习算法,利用其中的17种微量元素含量指标(Mn、Sr、Y、La、Ce、Pr、Nd、Sm、Eu、Gd、Tb、Dy、Ho、Er、Tm、Yb、Lu)和8种元素综合指标[(La/Nd)_(N)、(La/Yb)_(N)、(Gd/Yb)_(N)、Yb N、LREE、REE+Y、Eu/Eu^(*)、Sr/Y]建立了花岗岩的成因类型判别模型。结果表明随机森林(95.56%)和支持向量机(94.76%)的分类准确率均很高。发现Sr在区分I型和S型花岗岩方面最为重要,Ce和Mn在区分A型花岗岩方面最为重要,Eu对于区分S型及I/A型花岗岩方面较为重要。本文提出的机器学习模型便捷高效,可以快速地确定花岗岩的成因类型。
Granite is widely distributed in the earth,and different types of granite record different tectonic environments,source rock types,or deep-seated magmatic processes.Therefore,correctly identifying the genetic types of granites(I-type,S-type,A-type granites)is of great significance.Apatite,as one of the most common accessory minerals in granite,is rich in various major and trace elements,recording the source area and magmatic physicochemical characteristics of granite.Therefore,it has become an important indicator for distinguishing the genetic types of granites.This article collected trace element data of apatite from I-type,S-type,and A-type granites around the world.Two machine learning algorithms,Random Forest(RF)and Support Vector Machine(SVM),were employed to establish a discrimination model for granite genesis types using 17 trace element content indicators(Mn,Sr,Y,La,Ce,Pr,Nd,Sm,Eu,Gd,Tb,Dy,Ho,Er,Tm,Yb,Lu)and 8 comprehensive element indicators[(La/Nd)_(N),(La/Yb)_(N),(Gd/Yb)_(N),Yb N,LREE,REE+Y,Eu/Eu^(*),Sr/Y].The results indicate that both RF(95.56%)and SVM(94.76%)achieved high classification accuracies.Sr is the most significant in distinguishing I-type and S-type granites,Ce and Mn are the most significant in distinguishing A-type granites,and Eu is relatively important in distinguishing S-type from I/A-type granites.The machine learning models proposed in this study are convenient and efficient,allowing for the rapid determination of granite genesis types.
作者
韩凤歌
冷成彪
陈加杰
邹少浩
王大钊
HAN FengGe;LENG ChengBiao;CHEN JiaJie;ZOU ShaoHao;WANG DaZhao(National Key Laboratory of Uranium Resources Exploration-Mining and Nuclear Remote Sensing,East China University of Technology,Nanchang 330013,China;State Key Laboratory of Nuclear Resources and Environment,East China University of Technology,Nanchang 330013,China;School of Sciences,East China University of Technology,Nanchang 330013,China;School of Earth Sciences,East China University of Technology,Nanchang 330013,China)
出处
《岩石学报》
北大核心
2025年第2期737-750,共14页
Acta Petrologica Sinica
基金
第二次青藏高原综合科学考察研究项目(2021QZKK0301)
江西省国家级高层次人才创新创业项目(K20230004)联合资助.
