基于GEP-逻辑回归的地质异常信息分类预测——以东天山地区化探数据为例  

Classification and forecasting of geological anomaly based on GEP-logistic regression:a case study from geochemical data of eastern Tianshan

作　　者：桂州 陈建国[2,3] 王成彬[2,3] GUI Zhou;CHEN Jian-guo;WANG Cheng-bin(School of Geosciences,China University of Petroleum(East China),Qingdao 257061,China;State Key Laboratory of Geological Processes and Mineral Resources,China University of Geosciences,Wuhan 430074,China;Faculty of Earth Resources,China University of Geosciences,Wuhan 430074,China)

机构地区：[1]中国石油大学(华东)地球科学与技术学院,山东青岛257061 [2]中国地质大学(武汉)地质过程与矿产资源国家重点实验室,武汉430074 [3]中国地质大学(武汉)资源学院,武汉430074

出　　处：《桂林理工大学学报》2018年第1期34-40,共7页Journal of Guilin University of Technology

基　　金：国家科技支撑计划项目(2011BAB06B08-2);国家自然科学基金项目(41272361);中国地质调查局项目(1212011120986)

摘　　要：高效精准的识别地质异常对成矿预测至关重要。本文将GEP引入到逻辑回归方法中,并对传统单一模型进行改进,克服了单一模型的局限性和缺点。将模型应用于东天山地区的化探数据处理,对斑岩型-矽卡岩型成矿具有指示意义的元素进行了识别应用,结果表明:GEP-逻辑回归法在识别异常信息方面效果良好;与单一模型(神经网络、Knn、逻辑回归)验算结果对比,GEP-逻辑回归法表现更优。Efficient and accurate identification of geological anomalies is crucial for metallogenic prediction.In this paper,GEP is introduced into the logistic regression method,and the traditional single model is improved to overcome the limitations and disadvantages of the single model.The model was applied to geochemical data processing in the eastern Tianshan area,and it was identified and applied to elements of porphyry-skarn type mineralization.The results show that GEP-logical logistic regression method is effective in identifying abnormal information.Compared with the single model(neural network,Knn,logistic regression)check results,GEP-logical logistic regression performed better.

关 键 词：GEP 逻辑回归 地质异常 分类预测 化探数据 

分 类 号：P628[天文地球—地质矿产勘探]