[1]潘用科,贺紫平,夏克文,等.改进的协同训练半监督SVM在油层识别中的应用[J].郑州大学学报(工学版),2022,43(01):14-19.[doi:10.13705/j.issn.1671-6833.2022.01.001]
　PAN Yongke,HE Ziping,XIA Kewen,et al.Improved Co-training Semi-supervised SVM and Its Application in Oil Layer Recognition[J].Journal of Zhengzhou University (Engineering Science),2022,43(01):14-19.[doi:10.13705/j.issn.1671-6833.2022.01.001]
点击复制

改进的协同训练半监督SVM在油层识别中的应用()
分享到：

《郑州大学学报(工学版)》[ISSN:1671-6833/CN:41-1339/T]

卷:
43卷
期数:
2022年01期
页码:
14-19
栏目:
出版日期:
2022-01-09

文章信息/Info

Title:
Improved Co-training Semi-supervised SVM and Its Application in Oil Layer Recognition
作者:
潘用科; 贺紫平; 夏克文; 牛文佳
河北工业大学电子信息工程学院;

Author(s):
PAN Yongke;  HE Ziping;  XIA Kewen;  NIU Wenjia
School of Electronics and Information Engineering, Hebei University of Technology, Tianjin 300401, China
关键词:
Keywords:
semi-supervised support vector machine;  co-training;  QPSO;  data editing;  oil layer recognition
分类号:
TP18
DOI:
10.13705/j.issn.1671-6833.2022.01.001
文献标志码:
A
摘要:
实际石油测井中有标签数据获取代价昂贵，且大量低廉的无标签数据未被使用，如何利用有限的有标签样本及大量的无标签样本获取准确的油层分布有待解决。半监督学习方法同时利用少量有标签样本及大量无标签样本便可获取良好的分类模型。因此，基于半监督支持向量机（Semi-supervised support vector machine, S3VM），提出一种改进的基于量子行为粒子群优化(QPSO)的协同训练S3VM油层识别算法（QPSO-CS3VM）。首先，引入多视图的协同训练策略思想，构造两个独立的初始分类器，然后互相交换并标记无标签样本来提高总体油层识别精度。其次，为提高初始分类精度，引入了量子行为粒子群算法用以优化S3VM。然后，针对S3VM中错分类的无标签样本在训练中迭代导致错误累计，模型性能下降的问题，一种新颖的K近邻数据剪辑方法被用来预测无标签样本伪标签的置信度，从而减少因为传播的错分样本导致的模型性能恶化下降。最后，将该改进算法用于测井数据挖掘中的油层识别。经过对某油田两口井的实际测井资料进行处理，结果表明：本文提出的新颖半监督协同训练油层识别方法要优于传统的半监督分类算法，识别率更高，分类效果更显著，与全监督SVM算法相比较，得到相差不大的分类精度的同时，速度更快。
Abstract:
It is expensive to obtain labeled data in actual oil logging, and a large amount of cheap unlabeled data are not used. How to use limited labeled samples and a large number of unlabeled samples to obtain accurate oil layer distribution remains to be solved. The semi-supervised learning methods were widely used because they could obtain good classification models using both a small number of labeled samples and a large number of unlabeled samples. Therefore, based on a semi-supervised support vector machine (S3VM), an improved semi-supervised support vector machine based on co-training and quantum-behaved particle swarm optimization algorithm (QPSO-CS3VM) was proposed for oil layer recognition. Firstly, the multi-view-based co-training strategy combined with S3VM was used to construct two independent initial classifiers, and then exchanged and labelled unlabeled samples to improve the overall oil layer recognition accuracy. Secondly, in order to improve the initial classification accuracy of original classifiers, the quantum behavioral particle swarm algorithm was introduced to optimize S3VM. Finally, a newly nearest neighbor data editing approach was used to predict the confidence of the pseudo-labelling of unlabeled data to reduce the deterioration of model perfor-mance caused by misclassification of data. The improved co-training semi-supervised SVM proposed in this paper improved the classification accuracy by 5.00% and 3.12% compared to the traditional co-training algorithm by performing oil layer recognition on the logging data of the two wells. The algorithm proposed in this paper had high accuracy in oil layer recognition and had practical application.

参考文献/References:

［1］ ZHU C，XIA K，ZU B，et al．Multiple-kernel relevance vector machine based on SOCP［J］．Journal of computational information systems，2014，10 ( 24 ) : 10831 －10838． 

［2］ LIU L，XIA X，XIA K，et al． Oil layer recognition by support vector machine based on semi-definite programming［J］．Journal of computational information systems，2014，10( 6) : 2579－2586． 
［3］ ZHANG J N，XIA K W，HE Z P，et al．Dynamic multiswarm differential learning quantum bird swarm algorithm and its application in random forest classification model ［ J ］． Computational intelligence and neuroscience，2020，2020: 1－24． 
［4］ 韩嵩，韩秋弘．半监督学习研究的述评［J］．计算机工 程与应用，2020，56( 6) : 19－27． 
［5］ BLUM A，MITCHELL T．Combining labeled and unlabeled data with co-training［C］/ /Proceedings of the 1998 11th Annual Conference on Computational Learning Theory．New York: ACM，1998: 92－100． 
［6］ BIE T D，CＲISTIANINI N． Convex methods for transduction［J］． Advances in neural information processing systems，2004，16: 1－8． 
［7］ XU P C，LU W K，WANG B F． A semi-supervised learning framework for gas chimney detection based on sparse autoencoder and TSVM ［J］． Journal of geophysics and engineering，2019，16( 1) : 52－61． 
［8］ DAI D，LI H X，JIA X Y，et al．A co-training approach for sequential three-way decisions［J］． International journal of machine learning and cybernetics，2020，11 ( 5) : 1129－1139． 
［9］ TAＲIQ S，AKHTAＲ N，AFZAL H，et al． A novel co-training-based approach for the classification of mental illnesses using social media posts［J］． IEEE access，2019，7: 166165－166172． 
［10］ SUN J，XU W B，FENG B．A global search strategy of quantum-behaved particle swarm optimization［C］/ / IEEE Conference on Cybernetics and Intelligent Systems，2004．Piscataway: IEEE，2004: 111－116． 
［11］ ZHANG M L，ZHOU Z H． CoTrade: confident co-training with data editing［J］．IEEE transactions on systems，man，and cybernetics，part B ( cybernetics) ， 2011，41( 6) : 1612－1626．
 ［12］ 高岳林，武少华．基于自适应粒子群算法的机器人 路径规划［J］． 郑 州 大 学 学 报( 工 学 版) ，2020，41 ( 4) : 46－51.
 ［13］ 章昕，张飞，肖雄，等．基于量子粒子群算法-支持向 量机的 冷 连 轧 断 带 故 障诊 断［J］． 冶 金 自 动 化， 2020，44( 6) : 17－24． 
［14］ LUO J，WANG X Y，XU Y G．Vibration fault diagnosis for hydroelectric generating unit based on generalized S-transform and QPSO-SVM［C］/ /2019 IEEE Sustainable Power and Energy Conference ( iSPEC) ． Piscataway: IEEE，2019: 2133－2137． 
［15］ LIU L．Oil layer recognition by support vector machine based on quantum-behaved particle swarm optimization ［J］．Journal of information and computational science， 2014，11( 5) : 1511－1518． 
［16］ 周志华．基于分歧的半监督学习［J］．自动化学报， 2013，39( 11) : 1871－1878． 
［17］ 鲜焱，吕佳．基于核均值漂移聚类的改进局部协同 训练算法［J］．重庆师范大学学报( 自然科学版) ， 2020，37( 4) : 106－113． 
［18］ BAI J C，XIA K W，LIN Y L，et al．Attribute reduction based on consistent covering rough set and its application［J］．Complexity，2017，2017: 1－9．

更新日期/Last Update: 2022-01-09