Matrix factorization for multivariate time series analysis
arXiv:1903.05589 · doi:10.1214/19-EJS1630
Abstract
Matrix factorization is a powerful data analysis tool. It has been used in multivariate time series analysis, leading to the decomposition of the series in a small set of latent factors. However, little is known on the statistical performances of matrix factorization for time series. In this paper, we extend the results known for matrix estimation in the i.i.d setting to time series. Moreover, we prove that when the series exhibit some additional structure like periodicity or smoothness, it is possible to improve on the classical rates of convergence.
arXiv comment16 pages
References in corpus (14)
- User-friendly tail bounds for sums of random matrices
- Recovering low-rank matrices from few coefficients in any basis
- ROP: Matrix recovery via rank-one projections
- Matrix Completion With Noise
- Rank-based model selection for multiple ions quantum tomography
- Bayesian methods for low-rank matrix estimation: short survey and theoretical study
- Sparsity considerations for dependent observations
- Optimal large-scale quantum state tomography with Pauli measurements
- A Bayesian Approach for Noisy Matrix Completion: Optimal Rate under General Sampling Distribution
- Pseudo-Bayesian Quantum Tomography with Rank-adaptation
- An Oracle Inequality for Quasi-Bayesian Non-Negative Matrix Factorization
- Link Prediction in Graphs with Autoregressive Features
- Recovering Multiple Nonnegative Time Series From a Few Temporal Aggregates
- Online Forecasting Matrix Factorization