Research map / Learning theory and algorithms / Matrix and tensor methods
Integrated Principal Components Analysis
Tiffany M. Tang, Genevera I. Allen
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 |
|---|---|---|---|
| use Frobenius iPCA estimators instead of concatenated PCA, JIVE, MFA, distributed PCA | joint subspace estimation · Simulated three coupled matrices (n=150) | subspace recovery error | Improved |
| use multiplicative or additive Frobenius penalty instead of L1 penalty | iPCA · Simulated three coupled matrices (n=150) (dense Sigma) | subspace recovery error | Improved |
| add Laplacian noise to data (departure from normality) | Frobenius iPCA · Simulated three coupled matrices (n=150) with Laplacian error | subspace recovery error | No clear effect |
| simulate from JIVE model instead of Kronecker model | Frobenius iPCA · JIVE-model simulation, high noise | subspace recovery error | No clear effect |
| use Frobenius iPCA instead of JIVE with known ranks | JIVE · JIVE-model simulation, low noise | subspace recovery error | Improved |
| use MFA normalization instead of plain concatenated PCA | concatenated PCA · Simulated three coupled matrices (n=150) | subspace recovery error | Improved |
| increase joint signal strength | all methods · Simulated three coupled matrices (n=150) | subspace recovery error | Improved |
| increase number of samples | all methods · Simulated three coupled matrices (n=150) | subspace recovery error | Hurt |
| use multiplicative Frobenius iPCs as random forest predictors instead of other methods | ROSMAP analysis · ROSMAP (507 subjects) | random forest test error | Improved |
| use joint iPCs instead of combining PCs from three individual PCAs | ROSMAP analysis · ROSMAP (507 subjects) | random forest test error | Improved |
| use multiplicative Frobenius instead of additive L1 estimators on sparse Sigma | iPCA · sparse simulation | subspace recovery error | Hurt |
Closest papers by meaning
- Generalized probabilistic principal component analysis of correlated data (JMLR 2020 · similarity 0.86)
- Shared Independent Component Analysis for Multi-Subject Neuroimaging (NeurIPS 2021 · similarity 0.86)
- MFPCA: Multiscale Functional Principal Component Analysis (AAAI 2019 · similarity 0.86)
- Probabilistic Geometric Principal Component Analysis with application to neural data (ICLR 2025 · similarity 0.84)
- Principle Component Trees and Their Persistent Homology (AAAI 2024 · similarity 0.84)
- Disentangling Interpretable Factors with Supervised Independent Subspace Principal Component Analysis (NeurIPS 2024 · similarity 0.84)
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.
