| 基于机器学习的锆石成因分类研究 |
| |
| 引用本文： | 朱紫怡,周飞,王瑀,周统,侯照亮,邱昆峰.基于机器学习的锆石成因分类研究[J].地学前缘,2022,29(5):464-475. |
| |
| 作者姓名： | 朱紫怡 周飞 王瑀 周统 侯照亮 邱昆峰 |
| |
| 作者单位： | 1.中国地质大学(北京) 地球科学与资源学院, 北京 1000832.奥地利维也纳大学地质系, 维也纳 10903.中国地质大学 地质过程与矿产资源国家重点实验室, 北京 100083 |
| |
| 基金项目： | 国家自然科学基金项目(91962106);国家自然科学基金项目(42111530124);国家自然科学基金项目(42072087);国家重点研发计划项目(2019YFA0708603);中国高校产学研创新基金资助课题(2021ALA01006);111计划2.0(BP0719021);中国地质大学(北京)创新创业训练计划项目(S202111415004) |
| |
| 摘 要： | 锆石是在自然界中多种温压条件下能够稳定保存,并记录原岩年龄信息的副矿物。锆石微量元素能完整记录地质演化过程信息。通过微量元素分析锆石成因的研究已久,通常利用Th-U图解和LaN-(Sm/La)N图解等二元图解对锆石进行分类研究。然而,随着锆石研究的深入,以及二元图解无法呈现数据高维度信息的局限性,传统图解已经不能满足对锆石类型进行准确判别,且对已知类型的锆石出现判定偏差。因此,本文将地质大数据与机器学习相结合,训练出高维度锆石成因分类器。文中收集了3 498条不同成因类型的锆石微量元素数据,并通过测试和运用随机森林、支持向量机、人工神经网络和k近邻等4种机器学习算法,最终得出准确率为86.8%的线性支持向量机锆石成因分类器,用于锆石类型的判定与预测。这项工作为锆石分类研究提供了更高维度的判别手段,极大提高了微量元素分析成因结果的精度。将锆石微量元素数据与机器学习方法相结合,是大数据分析与机器学习技术在地球化学研究中的积极探索。
|
| 关 键 词： | 锆石 微量元素 成因 大数据分析 机器学习 |
| 收稿时间： | 2021-12-16 |
Machine learning-based approach for zircon classification and genesis determination |
| |
| ZHU Ziyi,ZHOU Fei,WANG Yu,ZHOU Tong,HOU Zhaoliang,QIU Kunfeng.Machine learning-based approach for zircon classification and genesis determination[J].Earth Science Frontiers,2022,29(5):464-475. |
| |
| Authors: | ZHU Ziyi ZHOU Fei WANG Yu ZHOU Tong HOU Zhaoliang QIU Kunfeng |
| |
| Institution: | 1. School of Earth Sciences and Resources, China University of Geosciences (Beijing), Beijing 100083, China2. Department of Geology, University of Vienna, Vienna 1090, Austria3. State Key Laboratory of Geological Process and Mineral Resources, China University of Geosciences (Beijing), Beijing 100083, China |
| |
| Abstract: | Zircon, a stable paragenetic mineral in various geological environments, has been recognized as a great tool in the chronological study of primary rocks. Trace elements in zircons may reveal geological evolutionary processes, and have long been used in zircon classification and zircon formation studies by binary diagram method, such as Th-U and LaN-(Sm/La)N diagrams. However, with the massive increase of zircon research, the traditional binary diagrams are no longer adequate for a precise determination of zircon types because binary plots cannot display higher dimensional information and therefore can lead to erroneous interpretation of zircon data. To address this issue, we take a machine learning approach to analyzing 3 498 zircon trace element data for different zircon genetic types to obtain high-dimensional zircon classification diagrams. We tested four machine learning algorithms (random forest, support vector machine, artificial neural network, and k-nearest neighbor) and consider support vector machine, with an 86.8% accuracy in predicting zircon type and origin, can best contribute to zircon classification. In addition to the development of a high-dimensional zircon classification diagram, this work also greatly improves the accuracy of zircon genesis determination using trace elements, and demonstrates the applicability of modern data science technique in geochemical research. |
| |
| Keywords: | zircon trace elements formation big data analysis machine learning |
|
| 点击此处可从《地学前缘》浏览原始摘要信息 |
| 点击此处可从《地学前缘》下载免费的PDF全文 |
|
