基于支持向量机算法的岩浆硫化物矿床硫元素预测模型  

Prediction Model of Sulfur Element in Magmatic Sulfide Deposits Based on Support Vector Machine Algorithm

作　　者：李猛猛[1,2] 刘敬党 梁天意[2] 谭亮 王刚 朱玺 Li Mengmeng;Liu Jingdang;Liang Tianyi;Tan Liang;Wang Gang;Zhu Xi(Mining Institute,Liaoning Technical University,Fuxin 123000,Liaoning,China;Liaoning Chemical Geology Exploration Institute Co.,Ltd.,Jinzhou 121000,Liaoning,China;CCTEG Shenyang Engineering Company,Shenyang 110015,China)

机构地区：[1]辽宁工程技术大学矿业学院,辽宁阜新123000 [2]辽宁省化工地质勘查院有限责任公司,辽宁锦州121000 [3]中煤科工集团沈阳设计研究院有限公司,沈阳110015

出　　处：《吉林大学学报（地球科学版）》2022年第3期866-878,共13页Journal of Jilin University：Earth Science Edition

基　　金：中央地勘基金项目(2011D3092)。

摘　　要：硫元素在岩浆硫化物矿床的形成过程中具有十分重要的作用.本文以加拿大拉布拉多Voisey’s Bay镍铜硫化物矿床的钻孔数据为实验数据,利用基于径向基函数(radialbasisfunction,RBF)核函数的支持向量机(support vector machine,SVM)算法对实验数据中镍、铜元素质量分数与硫元素质量分数进行相关性分析,进而建立基于硫元素质量分数的预测模型.通过SVM计算出硫元素质量分数的预测值,并与硫元素原值进行拟合分析.根据拟合分析偏差曲线的计算结果,将实验数据划分成3个阶段,分别为:硫饱和阶段,硫元素质量分数为3.12×10^(-6)~20.80×10^(-6),方差小于1.44×10^(-7),分维值为0.35,李雅普诺夫指数λ>0,该阶段处于混沌状态;硫成矿阶段,硫元素质量分数为6.60×10^(-7)~17.80×10^(-9),方差小于1.37×10^(-9),分维值为0.60,λ>0,该阶段处于混沌状态;硫流失阶段,硫元素质量分数为2.00×10^(-8)~38.00×10^(-8),方差小于1.56×10^(-10),分维值为0.94,λ=0,该阶段处于稳定状态.进而说明岩浆硫化物矿床的原始岩浆中镍、铜元素质量分数和硫元素质量分数总体上并不是线性关系,也并非呈简单的非线性关系,而是呈分阶段非线性关系.Sulfur plays an important role in the formation of magmatic sulfide deposits.Based on the drill core data of the Voisey’s Bay nickel-copper sulfide deposit in Labrador,Canada,the correlation between nickel and copper mass fraction and sulfur mass fraction was analyzed by support vector machine(support vector machine,SVM)and kernel function based on radial basis function(radial basis function,RBF),and the prediction model of sulfur was established based on sulfur mass fraction.In the prediction model,nickel and copper mass fractions are used as independent variables,sulfur mass fraction is used as the dependent variable,the predicted value of sulfur mass fraction is calculated by SVM to fit the original value.According to the calculation results of the fitting curve,the sulfur evolution can be divided into three stages:In the sulfur saturation stage,the mass fraction of sulfur element is 3.12×10^(-6)-20.80×10^(-6),the variance is less than 1.44×10^(-7),the fractal dimension is 0.35,λ>0,and the stage is chaotic;In the mineralization stage,the mass fraction of sulfur element is 6.60×10^(-7)-17.80×10^(-7),the variance is less than 1.37×10^(-9),the fractal dimension is 0.60,λ>0,and the stage is chaotic;In the sulfur loss stage,the sulfur mass fraction is 2.00×10^(-8)-38.0.×10^(-8),the variance is less than 1.56×10^(-10),the fractal dimension is 0.94,λ=0,and the stage is stable.The results show that the mass fraction of nickel,copper and sulfur in the immiscible sulfur-saturated magma is not linearly or simply nonlinearly correlated,but staged nonlinearly.

关 键 词：岩浆硫化物矿床 硫饱和 支持向量机 非线性 岩浆混合作用 预测模型 

分 类 号：P612[天文地球—矿床学]