《计算机应用》唯一官方网站 ›› 2022, Vol. 42 ›› Issue (2): 646-654.DOI: 10.11772/j.issn.1001-9081.2021041023

• 前沿与综合应用 • 上一篇    

页岩气储层预测的多标签主动学习算法

汪敏1(), 冯婷婷1, 闵帆2, 唐洪明3, 闫建平3, 廖纪佳3   

  1. 1.西南石油大学 电气信息学院, 成都 610500
    2.西南石油大学 计算机科学学院, 成都 610500
    3.西南石油大学 地球科学与技术学院, 成都 610500
  • 收稿日期:2021-06-15 修回日期:2021-07-05 接受日期:2021-07-09 发布日期:2021-11-02 出版日期:2022-02-10
  • 通讯作者: 汪敏
  • 作者简介:汪敏(1980—),女,湖南邵阳人,教授,硕士,CCF会员,主要研究方向:数据挖掘、主动学习;
    冯婷婷(1994—),女,山东济宁人,硕士研究生,主要研究方向:主动学习;
    闵帆(1973—),男,重庆人,教授,博士,CCF会员,主要研究方向:粒计算、推荐系统、主动学习;
    唐洪明(1966—),男,四川武胜人,教授,博士,主要研究方向:开发地质学、非常规油气储层评价;
    闫建平(1980—),男,内蒙古凉城人,教授,博士,主要研究方向:测井地质学、岩石物理、非常规储层测井评价;
    廖纪佳(1983—),男,四川绵竹人,博士,主要研究方向:沉积学、储层保护。
  • 基金资助:
    国家自然科学基金资助项目(62006200);中国石油-西南石油大学创新联合体科技合作项目(2020CX020000)

Multi-label active learning algorithm for shale gas reservoir prediction

Min WANG1(), Tingting FENG1, Fan MIN2, Hongming TANG3, Jianping YAN3, Jijia LIAO3   

  1. 1.School of Electrical Engineering and Information,Southwest Petroleum University,Chengdu Sichuan 610500,China
    2.School of Computer Science,Southwest Petroleum University,Chengdu Sichuan 610500,China
    3.School of Geoscience and Technology,Southwest Petroleum University,Chengdu Sichuan 610500,China
  • Received:2021-06-15 Revised:2021-07-05 Accepted:2021-07-09 Online:2021-11-02 Published:2022-02-10
  • Contact: Min WANG
  • About author:WANG Min, born in 1980, M. S., professor. Her research interests include data mining, active learning.
    FENG Tingting, born in 1994, M, S. candidate. Her research interests include active learning.
    MIN Fan, born in 1973, Ph. D., professor. His research interests include granular computing, recommender system, active learning.
    TANG Hongming, born in 1966, Ph. D., professor. His research interests include development geology, evaluation of unconventional oil and gas reservoirs.
    YAN Jianping, born in 1980, Ph. D., professor. His research interests include logging geology, rock physics, evaluation of unconventional reservoir logging.
    LIAO Jijia, born in 1983, Ph. D. His research interests include sedimentology, reservoir protection.
  • Supported by:
    National Natural Science Foundation of China(62006200);CNPC-Southwest Petroleum University Innovation Consortium Science and Technology Cooperation Project(2020CX020000)

摘要:

针对页岩气储层数据获取困难、标签稀缺、标注成本高昂的问题,提出一种多标准主动查询的多标签学习(MAML)算法。首先,考虑样本的信息性和代表性来对样本进行初步处理;其次,加入包括属性差异性和标签丰富性的样本丰富性约束,在此基础上选择有价值的样本进行标签查询;最后,利用多标签学习算法来预测剩余样本的标签。通过11个Yahoo数据集上的实验,将MAML算法与流行的多标签学习算法和主动学习算法进行比较,验证了MAML算法的优越性。然后将实验扩展到4个真实的页岩气测井数据集。在该实验中,与多标签学习算法:基于K最近邻的多标签(ML-KNN)学习方法、多标签学习的反向传播(BP-MLL)算法、具有全局和局部标签相关性的多标签学习方法(GLOCAL)和通过查询信息性和代表性样本的主动学习(QUIRE)方法相比,MAML算法在页岩气储层综合品质预测精度均值上分别提升了45个百分点、68个百分点、68个百分点和51个百分点。实验结果充分验证了MAML算法在页岩气储层甜点预测领域的实用性和优越性。

关键词: 多标签学习, 主动学习, 多标准优化, 查询, 甜点预测

Abstract:

Concerning the problems of the difficulties in obtaining, the limitation of labels, and the high cost of labeling of shale gas reservoir data, a Multi-standard Active query Multi-label Learning (MAML) algorithm was proposed. First of all, with the consideration of the informativeness and representativeness of the samples, the preliminary processing was performed on the samples. Secondly, the sample richness constraints including attribute differences and label richness were added, on this basis, the valuable samples were selected and the labels of these samples were queried. Finally, a multi-label learning algorithm was used to predict the labels of the remaining samples. Through experiments on eleven Yahoo datasets, the MAML algorithm was compared with popular multi-label learning algorithms and active learning algorithms, and the superiority of the MAML algorithm was proved. Then, the experiments were extended to four real shale gas well logging datasets. In these experiments, compared with the multi-label learning algorithms: Multi-Label Multi-Label K-Nearest Neighbor (ML-KNN), BackPropagation for Multi-Label Learning (BP-MLL), multi-label learning with GLObal and loCAL label correlation (GLOCAL) and active learning by QUerying Informative and Representative Examples (QUIRE), the MAML algorithm improved the average prediction accuracy of comprehensive quality of shale gas reservoirs by 45 percentage points, 68 percentage points, 68 percentage points, and 51 percentage points, respectively. The practicability and superiority of the MAML algorithm in the prediction of shale gas reservoir sweet spots are fully proved by these experimental results.

Key words: multi-label learning, active learning, multi-standard optimization, query, sweet spot prediction

中图分类号: