Canonical correlation analysis using within-class coupling

0167-8655/$ see front matter 2010 Elsevier B.V. A doi:10.1016/j.patrec.2010.09.025 q The work of O. Kursun was supported by Scienti nation Unit of Istanbul University under the grant YA ⇑ Corresponding author. Tel.: +90 212 473 7070/17 E-mail addresses: [email protected] (O. Kurs Alpaydin), [email protected] (O.V. Favorov). Fisher’s linear discriminant analysis (LDA) is one of the most popular supervised linear dimensionality reduction methods. Unfortunately, LDA is not suitable for problems where the class labels are not available and only the spatial or temporal association of data samples is implicitly indicative of class membership. In this study, a new strategy for reducing LDA to Hotelling’s canonical correlation analysis (CCA) is proposed. CCA seeks prominently correlated projections between two views of data and it has been long known to be equivalent to LDA when the data features are used in one view and the class labels are used in the other view. The basic idea of the new equivalence between LDA and CCA, which we call withinclass coupling CCA (WCCCA), is to apply CCA to pairs of data samples that are most likely to belong to the same class. We prove the equivalence between LDA and such an application of CCA. With such an implicit representation of the class labels, WCCCA is applicable both to regular LDA problems and to problems in which only spatial and/or temporal continuity provides clues to the class labels. 2010 Elsevier B.V. All rights reserved.