测井识别岩性新方法——支持向量机方法  

A New Method for Logging Lithology Identification—SVM

作　　者：于代国[1] 孙建孟[1] 王焕增[2] 陈伟中[2] 李召成[1] 张振成[1] 

机构地区：[1]石油大学地球资源与信息学院,山东东营257061 [2]塔里木油田分公司,新疆库尔勒841000

出　　处：《大庆石油地质与开发》2005年第5期93-95,共3页Petroleum Geology & Oilfield Development in Daqing

摘　　要：将基于统计学习理论的支持向量机方法引入到测井资料处理和解释中。在测井岩性识别方面研究发现,支持向量机方法克服了神经网络的固有缺陷,提供了一种识别岩性的新方法。应用效果表明该方法具有适应性强、识别精度高的优点。In this paper the SVM, which is based on statistic studying theory, is induced to logging data processing and interpretation. The research on logging lithology identification suggests that SVM can overcome the fundamental disadvantages of neural network, and it is a new method to identify lithology. The application of this method proves that it has high suitability and identification accuracy.

关 键 词：岩性识别 测井曲线 支持向量机 

分 类 号：P631.3[天文地球—地质矿产勘探]