基于共同决策方向矢量的多源迁移及其快速学习方法

张景祥, 王士同

电子学报 ›› 2015, Vol. 43 ›› Issue (7) : 1349-1355.

PDF(1517 KB)
PDF(1517 KB)
电子学报 ›› 2015, Vol. 43 ›› Issue (7) : 1349-1355. DOI: 10.3969/j.issn.0372-2112.2015.07.015
学术论文

基于共同决策方向矢量的多源迁移及其快速学习方法

  • 张景祥1,2, 王士同1
作者信息 +

Common-Decision-Vector Based Multiple Source Transfer Learning Classification and Its Fast Learning Method

  • ZHANG Jing-xiang1,2, WANG Shi-tong1
Author information +
文章历史 +

摘要

多源迁移学习提取了多个相似领域之间有用信息,提高了学习效率,但存在计算核矩阵的空间和时间复杂度较高的问题.提出了一种多源迁移学习方法,该方法基于结构风险最小框架理论,以共同决策方向矢量为基准,将多个相似领域的决策方向矢量嵌入到支持向量机的训练过程中,提高了目标领域分类器的分类性能.并结合核心向量机理论提出了共同决策方向矢量核心向量机,实现对大样本数据集的快速分类学习.模拟和真实数据集实验表明了所提算法的有效性.

Abstract

Multiple source transfer learning (MSTL) has been obtaining more and more applications especially from several related source domains to help the learning task on target domain.However,multiple source transfer learning algorithms often deal with the corresponding quadratic programming problems which may suffer a big computational burden caused by the kernel matrix computation.In this paper,a novel common-decision-vector based multiple source transfer classification learning (CDV-MSTL) is proposed which doesn't depend on the intrinsic structure of data.This algorithm is based on the structural risk minimization principle and the SVM like framework,so it has good adaptability and better accuracy.Based on the theory of CVM,CDV-MSTL is extended to its CVM based version which can realize fast training for large scale data.Extensive experiments on synthetic and real-world datasets demonstrate the significant improvement in classification performance obtained by the proposed algorithm over existing MSTL algorithm.

关键词

共同决策矢量 / 多源迁移学习 / 分类 / 核心集向量机

Key words

common decision vector / multiple source transfer learning / classification / core vector machine

引用本文

导出引用
张景祥, 王士同. 基于共同决策方向矢量的多源迁移及其快速学习方法[J]. 电子学报, 2015, 43(7): 1349-1355. https://doi.org/10.3969/j.issn.0372-2112.2015.07.015
ZHANG Jing-xiang, WANG Shi-tong. Common-Decision-Vector Based Multiple Source Transfer Learning Classification and Its Fast Learning Method[J]. Acta Electronica Sinica, 2015, 43(7): 1349-1355. https://doi.org/10.3969/j.issn.0372-2112.2015.07.015
中图分类号: TP391.4   

参考文献

[1] Pan S J L,Kwok J T,Yang Q.Transfer learning via dimensionality reduction[A].Proceedings of the 23rd International Conference on Artificial Intelligence [C].California,USA,2008.677-682.
[2] Pan S J L,Ni X.Cross-domain sentiment classification via spectral feature alignment [A].Proceedings of the 19th International Conference on World Wide Web[C].New York:ACM,2010.751-760.
[3] Zhuang F Z,Luo P,Xiong H,Xiong Y H.Cross-domain learning from multiple sources:a concensus regularization perspective[J].IEEE Transactions on Knowledge and Data Engineering,2010,22(12):1664-1678.
[4] Sun S.Multi-view Laplacian support vector machines[A].Lecture Notes in Computer Science[C].Germany:Springer,2011.209-222.
[5] Xu Z,Sun S.Multi-view transfer learning with Adaboost[A].Proceedings of the 23rd International Conference Tools with Artificial Intelligence[C].New York,2011.399-402.
[6] Ling X,Dai W,Xue G R,et al.Spectral domain-transfer learning[A].Proceedings of the 14th International Conference on Knowledge Discovery and Data Mining[C].USA:ACM,2008.488-496.
[7] Gao J,Fan W,Sun Y,et al.Heterogeneous source consensus learning via decision propagation and negotiation[A].Proceedings of the 15rd International Conference on Knowledge Discovery and Data Mining[C].USA:ACM,2009.339-348.
[8] 蒋亦樟,邓赵红,王士同.ML型迁移学习模糊系统[J].自动化学报,2012,38(9):1393-1409. Jiang Yizhang,Deng Zhaohong,Wang Shitong.Mamdani-Larsen type transfer learning fuzzy system[J].Acta Automatica Sinica,2012,38(9):1393-1409.(in Chinese)
[9] 胡文军,王士同,邓赵红.适合大样本快速训练的最大夹角间隔核心集向量机[J].电子学报,39(5):1178-1184. Hu Wen-jun,WANG Shi-tong,DENG Zhao-hong.Maximum vector angular margin core vector machine suitable for fast training for large datasets[J].Acta Electronica Sinica,39(5):1178-1184.(in Chinese)
[10] Deng Z H,Fu-Lai Chung,Wang S T.FRSDE:fast reduced set density estimator using minimal enclosing ball approximation[J].Pattern Recognition,2008,(41):1363-1372.
[11] Tsang I,Kwok J,Cheung P.Core vector machines:fast SVM training on very large data sets[J].Journal of Machine Learning Research,2005,6:363-392.
[12] Tax D M J,Duin R P W.Support vector domain description[J].Pattern Recognition Letters,1999,20:1191-1199.
[13] Tax D M J,Duin R P W.Support vector data description[J].Machine Learning,2004,54:45-66.
[14] 邓乃杨,田英杰.数据挖掘的新方法—支持向量机[M].北京:科学出版杜,2004. Deng Nai-yang,Tian Ying-jie.New Method in Data Mining:Support Vector Machine[M].Beijing:Science Press,2004.(in Chinese).
[15] Tsang I,Kwok J,Cheung P.Core vector machines:fast SVM training on very large data sets[J].Journal of Machine Learning Research,2005,6:363- 392.
[16] Tsang I,wok J.Generalized core vector machines[J].IEEE Trans on Neural Networks,2006,17(5):1126-1139.
[17] Suykens J A K,Lukas L,et al.Least squares support vector machine classifiers:a large scale algorithm[A].Proceedings of the European Conference on Circuit Theory and Design[C].Stresa,Italy,1999.839-842.
[18] Dai W,Yang Q,Xue G,et al.Boosting for transfer learning[A].Proceedings of the 24th International Conference on Machine Learning[C].New York:ACM,2007.193-200.
[19] Yao Y,Doretto G.Boosting for transfer learning with multiple sources[A].Proceedings of the 24th International Conference on Computer Vision and Pattern Recognition[C].New York:ACM,2010.1855-1862.
[20] Ablavsky V H,Becker C J.Transfer learning by sharing support vectors[OL].http://infoscience.epfl.ch/record/181360.
[21] Evgeniou T,Pontil M.Regularized multi-task learning[A].Proceedings of the 10th ACM SIGKDD International Conference on Knowledge Discovery in Data Mining[C].New York:ACM,2004.109-117.
[22] Dai W,Xue G,Yang Q,et al.Co-clustering based classification for out-of-domain documents[A].Proceedings of the 13th ACM SIGKDD International Conference on Knowledge Discovery in Data Mining[C].New York:ACM,2007:210-219.
[23] http://wang.ist.psu.edu/docs/related.shtml

基金

国家自然科学基金 (No.61202311,No.61272210); 江南大学博士科研基金 (No.JUDCF13031); 江苏省普通高校研究生科研创新计划基金 (No.CXLX13-748)

PDF(1517 KB)

2616

Accesses

0

Citation

Detail

段落导航
相关文章

/