摘要
机器学习技术已开始应用于成矿有利区预测中,在铀矿地质领域有必要开展相关的研究。文章确定了基于机器学习的铀成矿有利区预测的主要流程,明确了砂岩型铀资源机器学习模型建立的方法,针对所建立铀资源样本集合,开展了基于机器学习的预测试验,实现了试验地区的铀成矿有利区预测,对未参与建模的样本进行测试,正确率达到80%以上。
Artificial intelligence technology represented by machine learning has been gradually applied to mineralization prediction,and the relevant research is also need to be carry out for uranium exploration.This article determined the main process of machine learning-based prediction of favorable uranium mineralization areas,and clarified the method for establishing sandstone-type uranium resource machine learning models.For the established uranium resource sample set,a machine learning-based prediction experiment was carried out to conduct the experiment.In the prediction of favorable areas of uranium mineralization,the accuracy rate of tests on samples that did not participate in the learning was over 80%.
作者
李瀚波
叶发旺
张川
李新春
淦清清
LI Hanbo;YE Fawang;ZHANG Chuan;LI Xinchun;GAN Qingqing(National Key Laboratory of Remote Sensing Information and Imagery Analyzing Technology,Beijing Research Institute of Uranium Geology,Beijing 100029,China)
出处
《铀矿地质》
CAS
CSCD
2022年第6期1219-1225,共7页
Uranium Geology
基金
核能开发项目“基于航空高光谱与伽玛能谱的铀矿勘查技术研究”(编号:[2021]88)资助。
关键词
样本集合
标签
机器学习
铀资源
sample set
label
machine learning
uranium resources
