Abstract Non-negative matrix factorization and its extensions were applied to various areas (i.e., dimensionality reduction, clustering, etc.). When the original data are corrupted by outliers noise, most of non-negative methods cannot achieve robust learn a subspace with binary codes. This paper puts forward semi-supervised method for learning, called RSNMF, image clustering. For better cluste...

A band Lanczos algorithm for the iterative computation of eigenvalues and eigenvectors of a large sparse symmetric matrix is described and tested on numerical examples. It starts with a p dimensional subspace, and computes an orthonormal basis for the Krylov spaces of A, generated from this starting subspace, in which A is represented by a 2p + 1 band matrix, whose eigenvalues can be computed. ...

Techniques of information retrieval extract relevant documents in a collection, in response to a user query. As is well-known [2] information retrieval techniques based on exact literal matching, i.e., on direct comparisons between the columns of the term-document matrix and the query, may be inaccurate due to common problems of word usage such as synonymy and polysemy. Latent Semantic Indexing...

Fast adaptive subspace estimation plays an increasingly important role in modern signal processing. It forms the key ingredient in many sensor array signal processing algorithms, system identification, and several recently derived blind signal separation and equalization algorithms. The generic subspace estimation problem in these applications might be stated as follows. Suppose that we are giv...

The Continuation of Invariant Subspaces (CIS) algorithm produces a smoothly varying basis for an invariant subspace R(s) of a parameter-dependent matrix A(s). In the case when A(s) is the Jacobian matrix for a system that comes from a spatial discretization of a partial differential equation, it will typically be large and sparse. Cl matcont is a user-friendly matlab package for the study of dy...

A novel approach for sub-band based multi-microphone speech dereverberation is presented. In recent contribution a method utilizing the null subspace of the spatial-temporal correlation matrix of the received signals (obtained by the generalized eigenvalue decomposition (GEVD) procedure). The desired acoustic transfer functions (ATF-s) are shown to be embedded in these generalized eigenvectors....

This paper is concerned with subspace acceleration techniques for computing the Crawford number, that is, the distance between zero and the numerical range of a matrix A. Our approach is based on an eigenvalue optimization characterization of the Crawford number. We establish local convergence of order 1 + √ 2 ≈ 2.4 for an existing subspace method applied to such and other eigenvalue optimizati...