基于随机森林算法的南岭东段离子吸附型稀土矿成矿预测

打开文本图片集

中图分类号：P612 文献标识码：A 文章编号：2097-5465（2025）04-0020-08

Metallogenic Prediction of Ion-adsorption Rare Earth Deposits in the Eastern Part of Nanling Based on Random Forest Algorithm

LI Ruodan，LIU Xinxing， ZHANG Juan，GUAN Ziqiong，WANG Kaidi，LIU Meili，WANG Lijie

Hebei GEO University，Shijiazhuang O50031，China

Abstract：Ion-adsorptionrareearthoresareacriticalsoureofstrategicmetalsinSouth China.Thispaperfocusesontheionadsorptinrareearthdepositsintheastenpartofanlingregion.Guidedbytheendogenc-exogenicmetallgenicteory，amultifactorcollaorative metallgenic prediction modelisconstructedbasedonArcGIS platformandrandom forestalgorithm.Ninekey ore-controlingvariableswereextractedbythestructuralbufferanalysis，thenormalizeddiferencevegetationindex（NDVI）， calculationof topographicandgeomorphicparameters，includingelevation，slope，surfacecutingdepth，elevationcoeficientof variationandtopographicrelief，anddelineationofthedisrbutionrangeofore-formingparentrocks.Thestudyareawasdivided using a1000m×1 000 m grid unit，generating a balanced data set containing 221 positive samples （mine point units）and an equal number of negative samples.The prediction accuracy of the model test set was 77.6% ，and the AUC value was O.94，showing strong generalizationabilityTepredictionesultsindicatesignificantspatialdiferetiationinmtalogenicprobabilitycrosstestudy area.Cosideringthegeologicalbackgroundandcurrentexplorationstatus，atotaloffivemetalogenicprospectiveareaswere delineated，clasifiedintoAandBlevels，including thrA-levelprospectiveareasandandtwoB-levelprospectingareas.

Keywords：ion-adsorption rare earth type deposit；random forest；metallogenic prediction

来稿日期：2025-03-26 DOI：10.13937/j. cnki. hbdzdxxb.2025.04.003

基金项目：自然资源部离子型稀土资源与环境重点实验室基金（2022IRERE105）；国家自然科学基金（41702352）；河北地质大学河北省战略性关键矿产协同创新中心开放项目（2024403084）

作者简介：李若丹（2000—），女，黑龙江哈尔滨人，硕士研究生，主要研究方向为离子吸附型稀土矿成矿机制及成矿预测。（剩余9891字）