Randomized Single-View Algorithms for Low-Rank Matrix ApproximationReport as inadecuate




Randomized Single-View Algorithms for Low-Rank Matrix Approximation - Download this document for free, or read online. Document in PDF available to download.

Under review, 2016

This paper develops a suite of algorithms for constructing low-rank approximations of an input matrix from a random linear image of the matrix, called a sketch. These methods can preserve structural properties of the input matrix, such as positive-semideniteness, and they can produce approximations with a user-specied rank. The algorithms are simple, accurate, numerically stable, and provably correct. Moreover, each method is accompanied by an informative error bound that allows users to select parameters a priori to achieve a given approximation quality. These claims are supported by computer experiments.

Keywords: dimension reduction ; matrix approximation ; numerical linear algebra ; randomized algorithm ; single-pass algorithm ; single-view algorithm ; streaming algorithm ; subspace embedding Reference EPFL-WORKING-221094





Author: Tropp, Joel Aaron; Yurtsever, Alp; Udell, Madeleine; Cevher, Volkan

Source: https://infoscience.epfl.ch/record/221094?ln=en







Related documents