Asymptotic Theory for Estimating the Singular Vectors and Values of a Partially-observed Low Rank Matrix with Noise
نویسندگان
چکیده
Matrix completion algorithms recover a low rank matrix from a small fraction of the entries, each entry contaminated with additive errors. In practice, the singular vectors and singular values of the low rank matrix play a pivotal role for statistical analyses and inferences. This paper proposes and studies estimators of these quantities. When the dimensions of the matrix increase to infinity and the probability of observing each entry is identical, Theorem 4.2 gives the asymptotic distribution of the estimated singular values. Even though the estimators use only a partially observed matrix, they achieve the same rates of convergence as the fully observed case. The ensuing estimator of the full low rank matrix is computed with a non-iterative algorithm and it provides a consistent estimator. In the cases studied in this paper, its convergence rate is comparable to the convergence rate of iterative matrix completion algorithms. The numerical experiments corroborate our theoretical results.
منابع مشابه
A Novel Noise Reduction Method Based on Subspace Division
This article presents a new subspace-based technique for reducing the noise of signals in time-series. In the proposed approach, the signal is initially represented as a data matrix. Then using Singular Value Decomposition (SVD), noisy data matrix is divided into signal subspace and noise subspace. In this subspace division, each derivative of the singular values with respect to rank order is u...
متن کاملA Novel Noise Reduction Method Based on Subspace Division
This article presents a new subspace-based technique for reducing the noise of signals in time-series. In the proposed approach, the signal is initially represented as a data matrix. Then using Singular Value Decomposition (SVD), noisy data matrix is divided into signal subspace and noise subspace. In this subspace division, each derivative of the singular values with respect to rank order is u...
متن کاملTHE OPTIMAL HARD THRESHOLD FOR SINGULAR VALUES IS 4/√3 By
We consider recovery of low-rank matrices from noisy data by hard thresholding of singular values, in which empirical singular values below a prescribed threshold λ are set to 0. We study the asymptotic MSE (AMSE) in a framework where the matrix size is large compared to the rank of the matrix to be recovered, and the signal-to-noise ratio of the low-rank piece stays constant. The AMSE-optimal ...
متن کاملGeneralized SURE for optimal shrinkage of singular values in low-rank matrix denoising
We consider the problem of estimating a low-rank signal matrix from noisy measurements under the assumption that the distribution of the data matrix belongs to an exponential family. In this setting, we derive generalized Stein’s unbiased risk estimation (SURE) formulas that hold for any spectral estimators which shrink or threshold the singular values of the data matrix. This leads to new data...
متن کاملThe singular values and vectors of low rank perturbations of large rectangular random matrices
In this paper, we consider the singular values and singular vectors of finite, low rank perturbations of large rectangular random matrices. Specifically, we prove almost sure convergence of the extreme singular values and appropriate projections of the corresponding singular vectors of the perturbed matrix. As in the prequel, where we considered the eigenvalues of Hermitian matrices, the non-ra...
متن کامل