Semi-supervised classification based on random subspace dimensionality reduction

• Published in: Pattern recognition Vol. 45; no. 3; pp. 1119 - 1135
• Main Authors: Yu, Guoxian, Zhang, Guoji, Domeniconi, Carlotta, Yu, Zhiwen, You, Jane
• Format: Journal Article
• Language: English
• Published: Kidlington Elsevier Ltd 01.03.2012; Elsevier
• Subjects: Applied sciences; Classification; Classifiers; Construction; Dimensionality reduction; Ensembles of classifiers; Exact sciences and technology; Face recognition; Graph construction; Graphs; Image processing; Information, signal and communications theory; Learning; Pattern recognition; Random subspaces; Reduction; Semi-supervised classification; Signal and communications theory; Signal processing; Signal representation. Spectral analysis; Signal, noise; Subspaces; Telecommunications and information theory; Dimensionality reduction; Semi-supervised classification; Graph construction; Random subspaces; Ensembles of classifiers; Biometrics; Performance evaluation; Automatic classification; Face recognition; Image processing; Subspace method; Pattern recognition; Signal classification; Supervised classification; Dimension reduction; Supervised learning; Telephone call; Open market; Automatic recognition
• ISSN: 0031-3203; 1873-5142
• DOI: 10.1016/j.patcog.2011.08.024

Graph structure is vital to graph based semi-supervised learning. However, the problem of constructing a graph that reflects the underlying data distribution has been seldom investigated in semi-supervised learning, especially for high dimensional data. In this paper, we focus on graph construction for semi-supervised learning and propose a novel method called Semi-Supervised Classification based on Random Subspace Dimensionality Reduction, SSC-RSDR in short. Different from traditional methods that perform graph-based dimensionality reduction and classification in the original space, SSC-RSDR performs these tasks in subspaces. More specifically, SSC-RSDR generates several random subspaces of the original space and applies graph-based semi-supervised dimensionality reduction in these random subspaces. It then constructs graphs in these processed random subspaces and trains semi-supervised classifiers on the graphs. Finally, it combines the resulting base classifiers into an ensemble classifier. Experimental results on face recognition tasks demonstrate that SSC-RSDR not only has superior recognition performance with respect to competitive methods, but also is robust against a wide range of values of input parameters. ► Base classifiers in simple random subspaces are diverse but not accurate. ► Dimensionality reduction is applied in random subspaces instead of original space. ► Base classifiers in the processed random subspaces are diverse and accurate. ► The combined classifier is accurate and robust to a wide range of input values. ► Ensemble classifiers based on random subspaces can be enhanced by this technique.