论文信息 - Intelligent Data Fusion Using Sparse Representations and Nonlinear Dimensionality Reduction

Intelligent Data Fusion Using Sparse Representations and Nonlinear Dimensionality Reduction

Abstract : We propose a new method for performing data fusion and subsequent classification in an information-efficient manner. We argue that an algorithm that can find sparse, low-dimensional representations of data is an excellent candidate for data fusion and classification. Two recent developments in signal processing are investigated: 1) The use of over-determined dictionaries (e.g., frames), and 2) the use of so-called nonlinear dimensionality reduction techniques.

[1] D. Donoho,et al. Hessian eigenmaps: Locally linear embedding techniques for high-dimensional data , 2003, Proceedings of the National Academy of Sciences of the United States of America.

[2] Chih-Jen Lin,et al. LIBSVM: A library for support vector machines , 2011, TIST.

[3] Michael Elad,et al. From Sparse Solutions of Systems of Equations to Sparse Modeling of Signals and Images , 2009, SIAM Rev..

[4] Stéphane Lafon,et al. Diffusion maps , 2006 .

[5] S T Roweis,et al. Nonlinear dimensionality reduction by locally linear embedding. , 2000, Science.

[6] Vladimir Vapnik,et al. Statistical learning theory , 1998 .

[7] M. Maggioni,et al. Manifold parametrizations by eigenfunctions of the Laplacian and heat kernels , 2008, Proceedings of the National Academy of Sciences.