Research map / Learning theory and algorithms / Matrix and tensor methods
Positive Definite Sparse Covariance Estimation via Dual Space Optimization
Fengpei Li, Wenfu Xia, Ziping Zhao
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| solve the dual problem with proximal gradient (DPGM) instead of primal PGM | PDSCE solver · Block covariance, d=1000, rho=0.04 | runtime (s): 519.14 -> 3.59 | Improved |
| solve the dual problem (DPGM) instead of primal PGM | PDSCE solver · Toeplitz covariance, d=1000, rho=0.04 | runtime (s): 1183.75 -> 7.01 | Improved |
| solve the dual problem (DPGM) instead of primal PGM | PDSCE solver · Banded covariance, d=1000, rho=0.12 | runtime (s): 361.77 -> 13.86 | Improved |
| solve the dual problem (DPGM) instead of primal PGM | PDSCE solver · Block covariance, d=1000, rho=0.12 | runtime (s): 80.77 -> 12.60 | Improved |
| use DPGM instead of block coordinate minimization BCM | PDSCE solver · Block covariance, d=1000, rho=0.12 | runtime (s): 6700.25 -> 12.60 | Improved |
| use DPGM instead of proximal Newton PNM | PDSCE solver · Toeplitz covariance, d=1000, rho=0.08 | runtime (s): 2981.49 -> 12.44 | Improved |
| use DPGM instead of PGM | PDSCE solver · S&P 500 daily returns, d=415, rho=0.01 | runtime (s): 46.72 -> 0.239 | Improved |
| use DPGM instead of PNM | PDSCE solver · S&P 500 daily returns, d=415, rho=0.01 | runtime (s): 597.70 -> 0.239 | Improved |
| use DPGM instead of PGM (iteration count) | PDSCE solver · Block covariance, d=1000, rho=0.04 | iterations: 242 -> 20 | Improved |
| use DPGM instead of PGM (iteration count) | PDSCE solver · Toeplitz covariance, d=1000, rho=0.04 | iterations: 561 -> 27 | Improved |
| use DPGM instead of BCM (iteration count) | PDSCE solver · Block covariance, d=1000, rho=0.04 | iterations: 34 -> 20 | Improved |
| use the sparse PD covariance estimator (L1, adaptive Lasso, MCP) instead of the sample covariance | global minimum variance portfolio · S&P 500 daily returns, d=415 | portfolio risk | Improved |
Closest papers by meaning
- Sparse PCA via Covariance Thresholding (JMLR 2016 · similarity 0.88)
- A Well-Conditioned and Sparse Estimation of Covariance and Inverse Covariance Matrices Using a Joint Penalty (JMLR 2016 · similarity 0.88)
- Dual Principal Component Pursuit for Robust Subspace Learning: Theory and Algorithms for a Holistic Approach (ICML 2021 · similarity 0.88)
- Robust Gaussian Covariance Estimation in Nearly-Matrix Multiplication Time (NeurIPS 2020 · similarity 0.88)
- Implicit Bias of Projected Subgradient Method Gives Provable Robust Recovery of Subspaces of Unknown Codimension (ICLR 2022 · similarity 0.87)
- Fast and Provable Algorithms for Sparse PCA with Improved Sample Complexity (ICML 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
