CSDL Home IEEE Transactions on Pattern Analysis & Machine Intelligence 2008 vol.30 Issue No.08 - August

Subscribe

Issue No.08 - August (2008 vol.30)

pp: 1490-1495

Yi Guo , University of New England, Armidale

Junbin Gao , Charles Sturt Universtiy, Bathurst

Paul W. Kwan , University of New England, Armidale

DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/TPAMI.2008.74

ABSTRACT

In most existing dimensionality reduction algorithms, the main objective is to preserve relational structure among objects of the input space in a low dimensional embedding space. This is achieved by minimizing the inconsistency between two similarity/dissimilarity measures, one for the input data and the other for the embedded data, via a separate matching objective function. Based on this idea, a new dimensionality reduction method called Twin Kernel Embedding (TKE) is proposed. TKE addresses the problem of visualizing non-vectorial data that is difficult for conventional methods in practice due to the lack of efficient vectorial representation. TKE solves this problem by minimizing the inconsistency between the similarity measures captured respectively by their kernel Gram matrices in the two spaces. In the implementation, by optimizing a nonlinear objective function using the gradient descent algorithm, a local minimum can be reached. The results obtained include both the optimal similarity preserving embedding and the appropriate values for the hyperparameters of the kernel. Experimental evaluation on real non-vectorial datasets confirmed the effectiveness of TKE. TKE can be applied to other types of data beyond those mentioned in this paper whenever suitable measures of similarity/dissimilarity can be defined on the input data.

INDEX TERMS

Machine learning, Clustering, Visualization

CITATION

Yi Guo, Junbin Gao, Paul W. Kwan, "Twin Kernel Embedding",

*IEEE Transactions on Pattern Analysis & Machine Intelligence*, vol.30, no. 8, pp. 1490-1495, August 2008, doi:10.1109/TPAMI.2008.74REFERENCES