基于Weka平台地球化学异常分类研究  

Research on geochemical anomaly classification based on Weka platform

作　　者：辛舒怡 张海洪 朱骏 宋鹏飞 赵宇航 XIN Shuyi;ZHANG Haihong;ZHU Jun;SONG Pengfei;ZHAO Yuhang(Changchun Institute of Technology,Changchun 130021,Jilin,China)

机构地区：[1]长春工程学院,吉林长春130021

出　　处：《吉林地质》2024年第1期64-73,共10页Jilin Geology

基　　金：长春工程学院大学生创新创业项目基于Weka平台的沉积物地球化学数据挖掘研究(S202111437057);吉林省自然科学基金(20220101178JC)。

摘　　要：地球化学数据是应用地球化学研究的重要组成部分,是勘查工作的基础成果,地质地球化学数据的机器学习算法在研究地球化学异常分类中具有重要参考。本文以省内某区土壤地球化学Au、Cu等13种元素数据为基础,基于Weka平台J48决策树算法,对目标研究区域元素地球化学异常进行分类,在监督条件下结果表明,经过训练的模型基础上,异常分类测试集准确度89.8%、错误率10.1%,模型对异常分类具有较高准确度。异常的面积、nap值、断裂、浓集中心、矿床数量是最有效参数。Geochemical data is an important part of applied geochemical research and the basic achievement of exploration.The machine learning algorithm of geological data has an important reference in the study of geochemical anomaly classification.In this paper,based on soil geochemical data of 13 elements,such as Au and Cu,in a certain area of the Province,and the J48 decision tree algorithm of Weka platform,the geochemical anomalies of the target research area are classified.The results showed that under supervised conditions,the accuracy of the trained test set is 89.8%and the error rate is 10.1%,with the trained model showing high accuracy in anomaly classification.The most effective parameters include abnormal area,nap value,faults,concentration centers,and number of mineral deposits.

关 键 词：大数据 机器学习 WEKA 地球化学 异常 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]