| 49 | 0 | 250 |
| 下载次数 | 被引频次 | 阅读次数 |
随着覆盖区与深部找矿对象日益复杂，传统统计学习以及基于欧氏空间假设的深度学习模型（如卷积神经网络，CNN）在表征非规则格网、多源异质与强空间相关的地学数据时面临明显瓶颈。图神经网络（GNN）以非欧几里得表示学习为核心，能够在图结构上进行特征传播与关系建模，为成矿预测提供了新的技术范式。系统综述了GNN的基本概念与关键机制，梳理了图卷积网络（GCN）、图注意力网络（GAT）和图自编码器（GAE）等代表性模型及其在成矿预测中的应用进展，重点阐明消息传递机制如何联合刻画地质单元的空间自相关、构造连通与成因逻辑关联，并总结其在不规则采样、多证据层融合、长程依赖表达与结果空间连贯性提升方面的优势。综合评述表明，GNN在多金属矿产智能预测中取得了良好的效果，并在缓解样本稀缺与类别不平衡、提升可解释性与知识注入能力方面具有独特潜力。最后，探讨了现有的挑战并展望了未来的研究方向，旨在为构建智能矿产勘查体系提供参考。
Abstract：As exploration targets in covered area and at depth become increasingly complex, conventional statistical learning and deep learning models based on Euclidean assumptions(e. g., convolutional neural networks, CNN) face clear bottlenecks in representing geoscientific big data characterized by irregular grids, multi-source heterogeneity, and strong spatial dependence. Graph neural networks(GNN),centered on non-Euclidean representation learning, enable feature propagation and relational modeling on graph structures, provide a new technical paradigm for mineral prospectivity mapping. This paper made a systematically reviews on the fundamental concepts and key mechanisms of GNN, and summarized representative models—such as graph convolutional networks(GCN), graph attention networks(GAT),and graph autoencoders(GAE) —together with their application progress in mineral prospectivity mapping. We emphasized how message-passing mechanisms jointly capture spatial autocorrelation, structural connectivity, and genetic-logical associations among geological units, and synthesized their advantages in handling irregular sampling, fusing multiple evidence layers, modeling long-range dependencies, and enhancing spatial coherence of prediction results. Overall, GNNs have achieved promising performance in intelligent prediction of polymetallic mineral resources, and demonstrated distinctive potential for the alleviating sample scarcity and class imbalance, while improving interpretability and knowledge injection. Finally,we discuss current challenges and outline future research directions,so as to provide references for building an intelligent mineral exploration system.
1赵鹏大.大数据时代数字找矿与定量评价[J].地质通报，2015,34(7):1255-1259.Zhao Pengda.Digital mineral exploration and quantitative evaluation in the big data age[J].Geological Bulletin of China,2015,34(7):1255-1259(in Chinese）.
2成秋明.面向人类智能与人工智能融合的矿产资源预测新范式[J].地学前缘，2025,32(4):1-19.Cheng Qiuming.A new paradigm for mineral resource prediction based on human intelligenceartificial intelligence integration[J].Earth Science Frontiers,2025,32(4):1-19(in Chinese）.
3肖克炎，李程，唐瑞，等.大数据智能预测评价[J].地学前缘，2025,32(4):20-37.Xiao Keyan,Li Cheng,Tang Rui,et al.Big data intelligent prediction and evaluation[J].Earth Science Frontiers,2025,32(4):20-37(in Chinese）.
4 Agterberg F P.Computer Programs for Mineral Exploration[J].Science,1989:4913.
5 Qiuming Cheng,Agterberg F P.Fuzzy weights of evidence method and its application in mineral potential mapping[J].Natural Resources Research,1999,8(1):27-35.
6 Agterberg P F,Bonham-Carter F G.Measuring the performance of mineral-potential maps[J].Natural Resources Research,2005,14(1):1-17.
7张道军.逻辑回归空间加权技术及其在矿产资源信息综合中的应用[D].北京：中国地质大学（北京），2015.Zhang Daojun.Spatially weighted logistic regression and its application in mineral resources information synthesis[D].Beijing:China University of Geosciences(Beijing),2015(in Chinese）.
8 Brown W M,Gedeon T D,Groves D I,et al.Artificial neural networks:a new method for mineral prospectivity mapping[J].Journal of the Geological Society of Australia,2000,47(4):757-770.
9 Harris D,Zurcher L,Stanley M,et al.A comparative analysis of favorability mappings by weights of evidence,probabilistic neural networks,discriminant analysis,and logistic regression[J].Natural Resources Research,2003,12(4):241-255.
10 Porwal A,Carranza M J E,Hale M.Artificial neural networks for mineral-potential mapping:a case study from Aravalli province,western India[J].Natural resources research,2003,12(3):155-171.
11 Abedi M,Norouzi G,Bahroudi A.Support vector machine for multi-classification of mineral prospectivity areas[J].Computers and Geosciences,2012,46:272-283.
12 Zuo R,Carranza M J E.Support vector machine:A tool for mapping mineral prospectivity[J].Computers and Geosciences,2010,37(12):1967-1975.
13韩创益，王恩德，夏建明，等.基于贝叶斯推理的LS-SVM矿产资源定量预测[J].东北大学学报（自然科学版），2017,38(11):1633-1636.Han Chuangyi,Wang Ende,Xia Jianming,et al.Mineral resource quantitative prediction based on LS-SVM combining with Bayesian inference[J].Journal of Northeastern University (Natural Science),2017,38(11):1633-1636(in Chinese）.
14 Rodriguez-Galiano,Chica-Olmo,Chica-Rivas.Predictive modelling of gold potential with the integration of multisource information based on random forest:a case study on the Rodalquilar area,southern Spain[J].International Journal of Geographical Information Science,2014,28(7):1336-1354.
15 Carranza M J E,Laborte G A.Random forest predictive modeling of mineral prospectivity with small number of prospects and data with missing values in Abra (Philippines)[J].Computers and Geosciences,2015,74:60-70.
16 Rodriguez-Galiano,Sanchez-Castillo M,Chica-Olmo M,et al.Machine learning predictive models for mineral prospectivity:an evaluation of neural networks,random forest,regression trees and support vector machines[J].Ore Geology Reviews,2015,71:804-818.
17刘艳鹏，朱立新，周永章.卷积神经网络及其在矿床找矿预测中的应用--以安徽省兆吉口铅锌矿床为例[J].岩石学报，2018,34(11):3217-3224.Liu Yanpeng,Zhu Lixin,Zhou Yongzhang.Application of Convolutional Neural Network in prospecting prediction of ore deposits:Taking the Zhaojikou PbZn ore deposit in Anhui Province as a case[J].Acta Petrologica Sinica,2018,34(11):3217-3224(in Chinese）.
18张士红.基于深度学习的四川会理“拉拉式”铜矿找矿预测研究[D].北京：中国地质大学（北京），2020.Zhang Shihong.Research on mineral exploration prediction of“Lala-type”copper deposit in Huili,Sichuan based on deep learning[D].Beijing:China University of Geosciences (Beijing),2020(in Chinese）.
19周永章，陈烁，张旗，等.大数据与数学地球科学研究进展--大数据与数学地球科学专题代序[J].岩石学报，2018,34(2):255-263.Zhou Yongzhang,Chen Shuo,Zhang Qi,et al.Advances and prospects of big data and mathematical geoscience[J].Acta Petrologica Sinica,2018,34(2):255-263(in Chinese）.
20左仁广.基于数据科学的矿产资源定量预测的理论与方法探索[J].地学前缘，2021,28(3):49-55.Zuo Renguang.Theory and method of quantitative mineral resource prediction based on data science[J].Earth Science Frontiers,2021,28(3):49-55(in Chinese）.
21张鑫.基于深度学习的成矿远景区预测研究[D].江门：五邑大学，2024.Zhang Xin.Research on prediction of mineralized prospective areas based on deep learning[D].Jiangmen:Wuyi University,2024(in Chinese）.
22王桂安，殷宗耀，余先川.基于遥感高光谱数据的人工智能地质填图与矿产预测--现状、挑战与展望[J].科学技术与工程，2025[2026-01-20].https：//link.cnki.net/urlid/11.4688.T.20250909.1753.002.Wang Gui’an,Yin Zongyao,Yu Xianchuan.Artificial intelligence geological mapping and mineral prediction based on remote sensing hyperspectral data:current situation,challenges and prospects[J/OL].Science Technology and Engineering,2025[2026-01-20](in Chinese）.
23 Jie Z,Ganqu C,Shengding H,et al.Graph neural networks:a review of methods and applications[J].AI Open,2020,1:57-81.
24 Ziwei Z,Peng C,Wenwu Z.Deep learning on graphs:a survey[J].IEEE Transactions on Knowledge and Data Engineering,2022,34(1):249-270.
25陈志行，王正海，卜浩坚，等.基于融合地质信息图卷积神经网络的高光谱伟晶岩锂铍识别[J/OL].地球科学，2025[2026-01-20].https：//link.cnki.net/urlid/42.1874.P.20251129.1416.008.Chen Zhixing,Wang Zhenghai,Bu Haojian,et al.Ageology-informed graph convolutional network for identifying lithium-and beryllium-rich pegmatites from hyperspectral imagery[J/OL].Earth Science Frontiers,2025[2026-01-20](in Chinese）.
26吴博，梁循，张树森，等.图神经网络前沿进展与应用[J].计算机学报，2022,45(1):35-68.Wu Bo,Liang Xun,Zhang Shusen,et al.Advances and applications in graph neural network[J].Chinese Journal of Computers,2022,45(1):35-68(in Chinese）.
27王健宗，孔令炜，黄章成，等.图神经网络综述[J].计算机工程，2021,47(4):1-12.Wang Jianzong,Kong Lingwei,Huang Zhangcheng,et al.Survey of graph neural network[J].Computer Engineering,2021,47(4):1-12(in Chinese）.
28侯磊，刘金环，于旭，等.图神经网络研究综述[J].计算机科学，2024,51(6):282-298.Hou Lei,Liu Jinhuan,Yu Xu,et al.Review of graph neural networks[J].Computer Science,2024,51(6):282-298(in Chinese）.
29 Wu Z,Pan S,Chen F,et al.A comprehensive survey on graph neural networks[J].Co RR,2019,abs/1901.00596.
30马帅，刘建伟，左信.图神经网络综述[J].计算机研究与发展，2022,59(1):47-80.Ma Shuai,Liu Jianwei,Zuo Xin.Survey on graph neural network research and development[J],2022,59(1):47-80(in Chinese）.
31徐冰冰，岑科廷，黄俊杰，等.图卷积神经网络综述[J].计算机学报，2020,43(5):755-780.Xu Bingbing,Cen Keting,Huang Junjie,et al.Asurvey on graph convolutional neural network[J].Chinese Journal of Computers,2020,43(5):755-780(in Chinese）.
32刘俊奇，涂文轩，祝恩.图卷积神经网络综述[J].计算机工程与科学，2023,45(8):1472-1481.Liu Junqi,Tu Wenxuan,Zhu En.Survey on graph convolutional neural network[J].Computer Engineering and Science,2023,45(8):1472-1481(in Chinese）.
33 Bruna J,Zaremba W,Szlam A,et al.Spectral networks and locally connected networks on graphs[J].Co RR,2013,abs/1312.6203.
34 Defferrard M,Bresson X,Vandergheynst P.Convolutional neural networks on graphs with fast localized spectral filtering[J].Co RR,2016,abs/1606.09375.
35 Kipf N T,Welling M.Semi-supervised classification with graph convolutional networks[J].Co RR,2016,abs/1609.02907.
36万升，杨健，宫辰.基于图神经网络的高光谱图像分类研究进展[J].电子学报，2023,51(6):1687-1709.Wan Sheng,Yang Jian,Gong Chen.Advances of hyperspectral image classification based on graph Neural Networks[J].Acta Electronica Sinica,2023,51(6):1687-1709(in Chinese）.
37李军，余龙，段依琳，等.基于图神经网络的高光谱遥感图像分类前沿进展[J].遥感学报，2025,29(6):1681-1704.Li Jun,Yu Long,Duan Yilin,et al.Advances in graph neural network-based hyperspectral remote sensing image classification[J].National Remote Sensing Bulletin,2025,29(6):1681-1704(in Chinese）.
38 VeličkovićP,Cucurull G,Casanova A,et al.Graph attention networks[J].ar Xiv:Machine Learning,2017.
39 Zhang J,Shi X,Xie J,et al.Gaan:Gated attention networks for learning on large and spatiotemporal graphs[J].ar Xiv:preprint ar Xiv:1803.07294,2018.
40 Lyu L,Cheng J,Peng N,et al.Auto-encoder based Graph convolutional networks for online financial antifraud[C].IEEE Conference on Computational Intelligence for Financial Engineering and Economics,2019:1-6.
41 Kipf T N,Welling M.Variational graph auto-encoders[J].ar Xiv:preprint ar Xiv:1611.07308,2016.
42 Gilmer J,Schoenholz S S,Riley P F,et al.Neural message passing for quantum chemistry[C]//International Conference on Machine Learning.Pmlr,2017:1263-1272.
43 Zuo R,Xiong Y,Wang J,et al.Deep learning and its application in geochemical mapping[J].EarthScience Reviews,2019,192:1-14.
44 Renguang Z,Ying X.Graph deep learning model for mapping mineral prospectivity[J].Mathematical Geosciences,2022,55(1):1-21.
45 Sihombing M F,Palin M R,Hughes S H,et al.Improved mineral prospectivity mapping using graph neural networks[J].Ore Geology Reviews,2024,172:106215.
46 Lou Y,Liu Y.Mineral prospectivity mapping based on a novel self-ensembling graph convolutional network[J].Mathematical Geosciences,2025,57(4):1-28.
47 Cao C,Wang X,Yang F,et al.Attention-driven graph convolutional neural networks for mineral prospectivity mapping[J].Ore Geology Reviews,2025,180:106554.
48 Rui W,Linfu X,Yongsheng L,et al.Mineral prospectivity prediction based on the dynamic relation model Atten-GCN:a case study of gold prospecting in the Yingfengjie area,Shaanxi province (northern China)[J].Ore Geology Reviews,2025,176:106399.
49 Ying X,Renguang Z.An Interpretable graph attention network for mineral prospectivity mapping[J].Mathematical Geosciences,2023,56(2):169-190.
50 Chen Y,Chen B,Shaylan A.Semi-supervised graph convolutional networks for integrating continuous and binary evidential layers for mineral exploration targeting[J].Ore Geology Reviews,2024,173:106260.
51 Ying X,Renguang Z,Gubin Z.The graph attention network and its post-HOC explanation for recognizing mineralization-related geochemical anomalies[J].Applied Geochemistry,2023,155.
52 Qingfeng G,Shuliang R,Lirong C,et al.Recognizing multivariate geochemical anomalies related to mineralization by using deep unsupervised graph learning[J].Natural Resources Research,2022,31(5):2225-2245.
53 Luyi S,Ying X,Renguang Z.A heterogeneous graph construction method for mineral prospectivity mapping[J].Natural Resources Research,2024,33(4):1365-1376.
54 Sheng S,Wang Y,Tian J,et al.Knowledge-data collaboration-driven mineral prospectivity prediction with graph attention networks[J].Minerals (2075-163X),2025,15(11）.
55 Yan Q,Zhao J,Xue L,et al.Mineral prospectivity mapping based on spatial feature classification with geological map knowledge graph embedding:case study of gold ore prediction at Wulonggou,Qinghai province (Western China)[J].Natural Resources Research,2024,33(6):2385-2406.
56 Zuo R.Key technology for intelligent mineral prospectivity mapping:challenges and solutions[J].Science China Earth Sciences,2025,68(9):1-16.
57 Yan Q,Pei Y,Xue L,et al.Mineral prospectivity mapping:an interpretable classifier combining catchment basin and knowledge graph embedding[J].Ore Geology Reviews,2025,184:106758.
58左仁广.智能矿产预测的技术挑战与解决方案[J].中国科学：地球科学，2025,55(9):3104-3119.Zuo Renguang.Key technology for intelligent mineral prospectivity mapping:Challenges and solutions[J].Scientia Sinica Terrae,2025,55(9):3104-3119(in Chinese）.
59 Julian D,Dietmar R M,Rohitash C.Predicting the emplacement of Cordilleran porphyry copper systems using a spatio-temporal machine learning model[J].Ore Geology Reviews,2021,137:104300.
60 Farahbakhsh E,Zahirovic S,Mcinnes B,et al.Machine learning-based spatio-temporal prospectivity modeling of porphyry systems in the new Guinea and Solomon islands region[J].Tectonics,2025,44 (3):e2024TC007190.
61李楠，尹世滔，柳炳利，等.知识数据联合驱动的可解释智能矿产预测研究--以四川可尔因矿集区为例[J].地学前缘，2025,32(4):60-77.Li Nan,Yin Shitao,Liu Bingli,et al.A knowledge-data driven interpretable intelligent mineral prediction:a case study of the Keeryin mineral concentration area,Sichuan province[J].Earth Science Frontiers,2025,32(4):60-77(in Chinese）.
62史清平，刘武生.利用可解释机器学习揭示复杂砂岩型铀矿的地球化学控制因素--以内蒙古巴音戈壁盆地塔木素铀矿床为例[J].世界核地质科学，2025,42(5):1106-1122.Shi Qingping,Liu Wusheng.Unraveling the geochemical controls of a complex sandstone-type uranium deposit using explainable machine learning:a case study from Tamusu deposit,Bayingebi basin[J].World Nuclear Geoscience,2025,42(5):1106-1122(in Chinese）.
63左仁广，成秋明，许莹，等.可解释性矿产预测人工智能模型[J].中国科学：地球科学，2024,54(9):2917-2928.Zuo Renguang,Cheng Qiuming,Xu Ying,et al.Explainable artificial intelligence models for mineral prospectivity mapping[J].Scientia Sinica Terrae,2024,54(9):2917-2928(in Chinese）.
64 Parsa M,Cumani R.Class Label Representativeness in machine learning-based mineral prospectivity mapping[J].Natural Resources Research,2025,34(4):1-25.
65翟裕生.试论矿床成因的基本模型[J].地学前缘，2014,21(1):1-8.Zhai Yusheng.A Preliminary discussion on fundamental model of metallogenic mechanism[J].Earth Science Frontiers,2014,21(1):1-8(in Chinese）.
66户佐安，邓锦程，韩金丽，等.图神经网络在交通预测中的应用综述[J].交通运输工程学报，2023,23(5):39-61.Hu Zuoan,Deng Jincheng,Han Jinli,et al.Review on application of graph neural network in traffic prediction[J].Journal of Traffic and Transportation Engineering,2023,23(5):39-61(in Chinese）.
67王成彬，王明果，王博，等.融合知识图谱的矿产资源定量预测[J].地学前缘，2024,31(4):26-36.Wang Chengbin,Wang Mingguo,Wang Bo,et al.Knowledge graph-infused quantitative mineral resource forecasting[J].Earth Science Frontiers,2024,31(4):26-36(in Chinese）.
68林秋怡，左仁广.迁移学习及其在固体地球科学中的应用[J].地质科技通报，2025,44(1):346-356.Lin Qiuyi,Zuo Renguang.Transfer learning and its application in solid earth geoscience[J].Bulletin of Geological Science and Technology,2025,44(1):346-356(in Chinese）.
69王永志，温世博，李博文，等.大模型驱动的矿产资源智能预测超级智能体构建方法探索[J].地学前缘，2025,32(4):38-45.Wang Yongzhi,Wen Shibo,LI Bowen,et al.Construction technology of super-agents for intelligent mineral resources prediction driven by large model[J].Earth Science Frontiers,2025,32(4):38-45(in Chinese）.
70尹锦宇，朱鹏飞，王宝令，等.铀矿勘查知识库设计与系统实现[J].世界核地质科学，2025,42(2):307-316.Yin Jinyu,Zhu Pengfei,Wang Baoling,et al.Design and system implementation of uranium exploration knowledge repository[J].World Nuclear Geoscience,2025,42(2):307-316(in Chinese).
基本信息:
中图分类号:P627
引用信息:
[1]马东来,叶发旺.图神经网络在矿产预测中的应用综述[J].世界核地质科学,2026,43(01):160-174.
基金信息:
核技术研发项目（编号：HNKF202309(36)）资助~~
2026-02-15
2026-02-15