ISSN: 1673-825X    Imprint: Chongqing University of Posts and Telecommunications Journal
Volume listings
  • Current volume
  •  
  • Search by keywords:
  • Search by 
  • Keywords
  • To
  • Journal archive
Related Links
分支定界半监督SVM在油层识别中的应用
Oil layer recognition by semi-supervised support vector machine based on branch and bound
DOI:10.3979/j.issn.1673-825X.2019.04.018
Received:January 09, 2018  Revised:March 01, 2019
中文关键词:油层识别  半监督学习  支持向量机  分支定界
英文关键词:oil layer recognition  semi-supervised learning  support vector machine  branch and bound
基金项目:河北省自然科学基金(E2016202341);河北省高等学校科学技术研究项目(BJ2014013)
Author NameAffiliationE-mail
HE Ziping School of Information Engineering, Hebei University of Technology, Tianjin 300401, P. R. China 295990667@qq.com 
XIA Kewen School of Information Engineering, Hebei University of Technology, Tianjin 300401, P. R. China kwxia@hebut.edu.cn 
PAN Yongke School of Information Engineering, Hebei University of Technology, Tianjin 300401, P. R. China 259234914@ qq.com 
WANG Li School of Information Engineering, Hebei University of Technology, Tianjin 300401, P. R. China qhdzywl@163.com 
Hits: 18
Download times: 12
中文摘要:
      为解决油层识别中存在的获得有标记数据的代价过高,有标记数据稀少的问题,提出一种新的基于分支定界的半监督支持向量机(branch and bound for semi-supervised support vector machine,BBS3VM)的油层识别方法。此方法主要将半监督学习(semi-supervised learning, SSL)和分支定界的思想引入到支持向量机(support vector machine, SVM)分类算法中。通过半监督学习的思想,使用大量未标记的样本来改善学习性能,利用分支定界算法提高半监督支持向量机(semi-supervised support vector machine,S3VM)算法的分类精度,将此改进算法应用于测井数据挖掘中的油层识别。经过对某油田的实际测井资料进行处理,实验结果表明,半监督油层识别方法要优于传统的S3VM分类算法,识别率更高,分类效果更显著,与全监督的SVM算法相比较,得到相差不大的分类精度的同时,速度更快。
英文摘要:
      To solve the high price and the lack of labeled data in oil layer recognition, a new oil layer recognition by semi-supervised support vector machine based on branch and bound is proposed. The algorithm introduces semi-supervised learning and branch and bound to SVM. Firstly, semi-supervised learning is used to improve learning performance by a large number of unlabeled data. Secondly, branch and bound is used to improve the classification accuracy of S3VM, and then the improved algorithm to oil layer recognition is applied. Result shows that the algorithm proposed in this paper is superior to the traditional S3VM and the supervised SVM. The experimental results show that the algorithm proposed in this paper has higher recognition rate and better classification effect than the traditional S3VM algorithm, and it’s also faster than SVM algorithm when the classification accuracy of the two is almost equal.
HTML    View PDF   View/Add Comment  Download reader