Semisupervised Generalized Discriminant Analysis

被引:27
|
作者
Zhang, Yu [1 ]
Yeung, Dit-Yan [1 ]
机构
[1] Hong Kong Univ Sci & Technol, Dept Comp Sci & Engn, Kowloon, Hong Kong, Peoples R China
来源
IEEE TRANSACTIONS ON NEURAL NETWORKS | 2011年 / 22卷 / 08期
关键词
Constrained concave-convex procedure; dimensionality reduction; generalized discriminant analysis; semisupervised learning; FRAMEWORK; LDA;
D O I
10.1109/TNN.2011.2156808
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Generalized discriminant analysis (GDA) is a commonly used method for dimensionality reduction. In its general form, it seeks a nonlinear projection that simultaneously maximizes the between-class dissimilarity and minimizes the within-class dissimilarity to increase class separability. In real-world applications where labeled data are scarce, GDA may not work very well. However, unlabeled data are often available in large quantities at very low cost. In this paper, we propose a novel GDA algorithm which is abbreviated as semisupervised generalized discriminant analysis (SSGDA). We utilize unlabeled data to maximize an optimality criterion of GDA and formulate the problem as an optimization problem that is solved using the constrained concave-convex procedure. The optimization procedure leads to estimation of the class labels for the unlabeled data. We propose a novel confidence measure and a method for selecting those unlabeled data points whose labels are estimated with high confidence. The selected unlabeled data can then be used to augment the original labeled dataset for performing GDA. We also propose a variant of SSGDA, called M-SSGDA, which adopts the manifold assumption to utilize the unlabeled data. Extensive experiments on many benchmark datasets demonstrate the effectiveness of our proposed methods.
引用
收藏
页码:1207 / 1217
页数:11
相关论文
共 50 条
  • [41] Modified Class-Incremental Generalized Discriminant Analysis
    He, Yunhui
    2009 INTERNATIONAL CONFERENCE ON COMPUTER ENGINEERING AND TECHNOLOGY, VOL I, PROCEEDINGS, 2009, : 262 - 266
  • [42] Generalized Consistent Error Estimator of Linear Discriminant Analysis
    Zollanvari, Amin
    Dougherty, Edward R.
    IEEE TRANSACTIONS ON SIGNAL PROCESSING, 2015, 63 (11) : 2804 - 2814
  • [43] Modified generalized discriminant analysis for radar HRRP recognition
    Hualin Liu
    Wanlin Yang
    2007 5TH INTERNATIONAL CONFERENCE ON MICROWAVE AND MILLIMETER WAVE TECHNOLOGY PROCEEDINGS, 2007, : 840 - +
  • [44] Semisupervised Classification for Industrial Process Faults Using Modified Fisher Discriminant Analysis Assisted by Stacked Autoencoder
    Deng, Xiaogang
    Dai, Jiabing
    Cao, Yuping
    2019 CHINESE AUTOMATION CONGRESS (CAC2019), 2019, : 2282 - 2287
  • [45] Generalized MDS for data exploration, discriminant analysis, clustering and visualization
    Johannsen, DA
    Solka, J
    PROCEEDINGS OF THE 8TH JOINT CONFERENCE ON INFORMATION SCIENCES, VOLS 1-3, 2005, : 1739 - 1742
  • [46] Characterization of a family of algorithms for generalized discriminant analysis on undersampled problems
    Ye, JP
    JOURNAL OF MACHINE LEARNING RESEARCH, 2005, 6 : 483 - 502
  • [47] Kernel generalized nonlinear discriminant analysis algorithm for pattern recognition
    Dai, G
    Qian, YM
    ICIP: 2004 INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, VOLS 1- 5, 2004, : 2697 - 2700
  • [48] An approach for mechanical fault classification based on generalized discriminant analysis
    Li W.-H.
    Shi T.-L.
    Yang S.-Z.
    Frontiers of Mechanical Engineering in China, 2006, 1 (3): : 292 - 298
  • [49] Generalized robust linear discriminant analysis for jointly sparse learning
    Zhu, Yufei
    Lai, Zhihui
    Gao, Can
    Kong, Heng
    APPLIED INTELLIGENCE, 2024, 54 (19) : 9508 - 9523
  • [50] Feature reduction via generalized uncorrelated linear discriminant analysis
    Ye, Jieping
    Janardan, Ravi
    Li, Qi
    Park, Haesun
    IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, 2006, 18 (10) : 1312 - 1322