pith. sign in

arxiv: 2606.23174 · v1 · pith:NBSHOZQEnew · submitted 2026-06-22 · 📊 stat.ME

Principal Covariate Regression with Nuclear Norm Penalty

Kaiwen Liu , Lisa Verbeij , Wouter Weeda , Mark de Rooij This is my paper

Pith reviewed 2026-06-26 07:32 UTC · model grok-4.3

classification 📊 stat.ME
keywords principal covariate regressionnuclear norm penaltydimensionality reductionregularized regressionhigh-dimensional datasimultaneous optimizationPCovR
0
0 comments X

The pith

A nuclear norm penalty lets principal covariate regression reduce dimensions and estimate regularized coefficients at the same time.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Existing principal covariate regression methods require researchers to decide the order of dimensionality reduction and coefficient estimation, often through separate steps and ad-hoc rules. The paper proposes PcovRnnp, which adds a nuclear norm penalty to the PCovR objective so both tasks occur jointly during a single optimization. This setup aims to remove the need for sequential choices or post-processing selections. A reader would care because high-dimensional data analysis often suffers from inconsistent results when steps are ordered manually. The method targets settings where both reduction and regularization must be balanced without extra tuning layers.

Core claim

The authors claim that incorporating a nuclear norm penalty into the principal covariate regression objective enables simultaneous dimension reduction and estimation of regularized coefficients, solving the ordering problem that forces existing PCovR methods to perform these tasks sequentially.

What carries the argument

The nuclear norm penalty added to the PCovR objective function, which induces low-rank structure to couple dimensionality selection with coefficient estimation in one optimization pass.

If this is right

  • Dimension selection and regularized coefficient estimation occur jointly without manual ordering.
  • No additional post-hoc rules are needed to choose the number of dimensions after regression.
  • The approach applies directly to high-dimensional settings where both reduction and prediction are required.
  • Researchers avoid reintroducing ordering artifacts when balancing the two goals.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The penalty might generalize to other joint reduction-regression problems where sequential tuning creates bias.
  • Performance comparisons on benchmark datasets could reveal whether joint optimization improves prediction accuracy over staged methods.
  • In applications like neuroimaging or genomics, the method could reduce sensitivity to arbitrary step ordering.

Load-bearing premise

Adding a nuclear norm penalty to the PCovR objective is enough to force simultaneous rather than sequential optimization without needing extra tuning parameters or post-hoc selection rules.

What would settle it

A dataset or simulation where PcovRnnp still requires separate dimension selection after fitting or produces results equivalent to sequential PCovR with arbitrary ordering would show the penalty does not achieve true simultaneity.

Figures

Figures reproduced from arXiv: 2606.23174 by Kaiwen Liu, Lisa Verbeij, Mark de Rooij, Wouter Weeda.

Figure 1
Figure 1. Figure 1: Root Mean Squared Error of Y in Simulation 1. Note. The legend denotes different methods and their hyperparameter selection rules. PcovRnnp(λ1se,Y/YX): PcovRnnp using the 1-standard-error rule (λ1se) targeting only Y or both Y and X as CV objectives. PcovRnnp(λmin,Y/YX): PcovRnnp using the minimum CV error rule (λmin) targeting only Y or both Y and X. Ridge(λ1se)/Ridge(λmin): Ridge regression using λ1se an… view at source ↗
Figure 2
Figure 2. Figure 2: Root Mean Squared Error of Coefficients in Simulation 1. Note. The legend denotes different methods and their hyperparameter selection rules. PcovRnnp(λ1se,Y/YX): PcovRnnp using the 1-standard-error rule (λ1se) targeting only Y or both Y and X as CV objectives. PcovRnnp(λmin,Y/YX): PcovRnnp using the minimum CV error rule (λmin) targeting only Y or both Y and X. Ridge(λ1se)/Ridge(λmin): Ridge regression us… view at source ↗
Figure 3
Figure 3. Figure 3: Number of Components Selected by PcovRnnp in Simulation 1. Note. The legend denotes different methods and their hyperparameter selection rules. PcovRnnp(λ1se,Y/YX): PcovRnnp using the 1-standard-error rule (λ1se) targeting only Y or both Y and X as CV objectives. PcovRnnp(λmin,Y/YX): PcovRnnp using the minimum CV error rule (λmin) targeting only Y or both Y and X. Test sample size Ntest = 1000. The horizon… view at source ↗
Figure 4
Figure 4. Figure 4: Root Mean Squared Error of Y in Simulation 2. Note. PcovRnnp: PcovRnnp using the 1-standard-error rule (λ1se) targeting only Y as CV objective. Ridge: Ridge regression using λ1se rules. Test sample size Ntest = 1000. Results are averaged over 100 repetitions [PITH_FULL_IMAGE:figures/full_fig_p016_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Root Mean Squared Error of Coefficients in Simulation 2. Note. PcovRnnp: PcovRnnp using the 1-standard-error rule (λ1se) targeting only Y as CV objective. Ridge: Ridge regression using λ1se rules. Test sample size Ntest = 1000. Results are averaged over 100 repetitions [PITH_FULL_IMAGE:figures/full_fig_p017_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Number of Components Selected by PcovRnnp in Simulation 2. Note. PcovRnnp: PcovRnnp using the 1-standard-error rule (λ1se) targeting only Y as CV objective. Ridge: Ridge regression using λ1se rules. Test sample size Ntest = 1000. The horizontal dashed line refers to True rank R = 4. The horizontal dotted line refers to total predictor rank R + S = 10, where the true predictive rank R = 4 and non-predictive… view at source ↗
read the original abstract

In high-dimensional data settings, dimensionality reduction or variable selection are key steps when using statistical learning techniques. Principal Covariate Regression-type methods aim to perform both dimensionality reduction and (regularized) regression steps in one analysis. However, existing PCovR methods cannot simultaneously select dimensionalities and estimate regularized coefficients, forcing researchers to make ad-hoc choices in the order of these steps. In this study, we propose a novel method called Principal Covariate Regression with Nuclear Norm Penalty (PcovRnnp) that allows simultaneous dimension reduction and estimation of regularized coefficients.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript proposes Principal Covariate Regression with Nuclear Norm Penalty (PcovRnnp), which augments the PCovR objective with a nuclear norm penalty to perform simultaneous dimensionality reduction and regularized coefficient estimation, thereby avoiding the ad-hoc ordering of steps required by existing PCovR methods in high-dimensional settings.

Significance. If the formulation achieves joint optimization without reintroducing separate selection steps, the method could streamline high-dimensional regression analyses by integrating dimension reduction and regularization in a single procedure.

major comments (2)
  1. [Abstract] Abstract: the claim that the nuclear norm penalty 'allows simultaneous dimension reduction and estimation of regularized coefficients' is not accompanied by any explicit objective function, optimization procedure, or derivation showing that the penalty enforces joint rather than sequential optimization.
  2. [Abstract] Abstract: the nuclear norm penalty is governed by a hyperparameter λ whose selection (via cross-validation or otherwise) is not addressed; this selection step risks recreating the dimensionality-ordering problem the method claims to solve, which is load-bearing for the simultaneity claim.
minor comments (1)
  1. The abstract would be strengthened by a one-sentence statement of the base PCovR objective being extended.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive comments. We address each major comment below, indicating revisions where the manuscript will be updated to improve clarity.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the claim that the nuclear norm penalty 'allows simultaneous dimension reduction and estimation of regularized coefficients' is not accompanied by any explicit objective function, optimization procedure, or derivation showing that the penalty enforces joint rather than sequential optimization.

    Authors: The abstract is a high-level summary; the explicit objective function appears in Equation (2) of Section 2, which augments the standard PCovR loss with the nuclear-norm term on the coefficient matrix. The optimization procedure (alternating minimization) and the derivation that the nuclear norm induces joint low-rank structure and coefficient estimation are given in Section 2.3 and the supplement. We will revise the abstract to include a brief reference to the objective and the joint character of the solution. revision: yes

  2. Referee: [Abstract] Abstract: the nuclear norm penalty is governed by a hyperparameter λ whose selection (via cross-validation or otherwise) is not addressed; this selection step risks recreating the dimensionality-ordering problem the method claims to solve, which is load-bearing for the simultaneity claim.

    Authors: λ is selected by cross-validation on out-of-sample prediction error. For any fixed λ the single optimization problem simultaneously yields both the effective rank (dimension reduction) and the regularized coefficients; no separate choice of dimensionality is required before or after regression. This is the key distinction from sequential PCovR pipelines. We will add a short subsection on hyperparameter selection to make this distinction explicit. revision: partial

Circularity Check

0 steps flagged

No significant circularity; direct method proposal without self-referential reduction

full rationale

The abstract and available text describe a straightforward proposal to augment the PCovR objective with a nuclear norm penalty for joint optimization. No equations, fitted parameters, or self-citations are shown that would make any claimed prediction or result equivalent to its inputs by construction. The simultaneity claim is presented as a consequence of the penalty formulation itself rather than a quantity defined in terms of its own outputs. This is a standard non-circular method-construction paper; score 0 is the expected outcome when no load-bearing derivation reduces to a fit or self-citation chain.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only; no free parameters, axioms, or invented entities are described.

pith-pipeline@v0.9.1-grok · 5617 in / 1080 out tokens · 21754 ms · 2026-06-26T07:32:55.760820+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

106 extracted references · 94 canonical work pages · 3 internal anchors

  1. [1]

    NeuroImage , author =

    The. NeuroImage , author =. 2012 , pages =. doi:10.1016/j.neuroimage.2012.02.018 , language =

    work page doi:10.1016/j.neuroimage.2012.02.018 2012

  2. [2]

    NeuroImage , author =

    The. NeuroImage , author =. 2013 , pages =. doi:10.1016/j.neuroimage.2013.05.041 , language =

    work page doi:10.1016/j.neuroimage.2013.05.041 2013

  3. [3]

    Neuroimage , author =

    Group comparison of resting-state. Neuroimage , author =. 2009 , pages =

    2009

  4. [4]

    Neuroimage , author =

    Function in the human connectome: task-. Neuroimage , author =. 2013 , pages =

    2013

  5. [5]

    Nature neuroscience , author =

    A positive-negative mode of population covariation links brain connectivity, demographics and behavior , volume =. Nature neuroscience , author =. 2015 , pages =

    2015

  6. [6]

    , month = jul, year =

    Helwig, Nathaniel E. , month = jul, year =. grpnet:. doi:10.32614/CRAN.package.grpnet , abstract =

    work page doi:10.32614/cran.package.grpnet

  7. [7]

    Information Fusion , author =

    Stacked penalized logistic regression for selecting views in multi-view learning , volume =. Information Fusion , author =. 2020 , pages =. doi:10.1016/j.inffus.2020.03.007 , language =

    work page doi:10.1016/j.inffus.2020.03.007 2020

  8. [8]

    Regularized

    Cotugno, Lorenza and Rooij, Mark de and Siciliano, Roberta , month = dec, year =. Regularized. doi:10.48550/arXiv.2511.16718 , abstract =

    work page doi:10.48550/arxiv.2511.16718

  9. [9]

    Journal of Statistical Software , author =

    Supervised. Journal of Statistical Software , author =. doi:10.18637/jss.v086.i01 , language =

    work page doi:10.18637/jss.v086.i01

  10. [10]

    Journal of Multivariate Analysis , author =

    Linear mixed models and penalized least squares , volume =. Journal of Multivariate Analysis , author =. 2004 , pages =. doi:10.1016/j.jmva.2004.04.013 , language =

    work page doi:10.1016/j.jmva.2004.04.013 2004

  11. [11]

    Bates, M

    Fitting. Journal of Statistical Software , author =. doi:10.18637/jss.v067.i01 , language =

    work page doi:10.18637/jss.v067.i01

  12. [12]

    The Annals of Applied Statistics , author =

    Pathwise coordinate optimization , volume =. The Annals of Applied Statistics , author =. doi:10.1214/07-AOAS131 , number =

    work page doi:10.1214/07-aoas131

  13. [13]

    Regularization paths for generalized linear models via coordinate descent

    Regularization. Journal of Statistical Software , author =. doi:10.18637/jss.v033.i01 , language =

    work page doi:10.18637/jss.v033.i01

  14. [14]

    Statistics and Computing , author =

    Group descent algorithms for nonconvex penalized linear and logistic regression models with grouped predictors , volume =. Statistics and Computing , author =. 2015 , pages =. doi:10.1007/s11222-013-9424-2 , language =

    work page doi:10.1007/s11222-013-9424-2 2015

  15. [15]

    Journal of Computational and Graphical Statistics , author =

    Versatile. Journal of Computational and Graphical Statistics , author =. 2025 , pages =. doi:10.1080/10618600.2024.2362232 , language =

    work page doi:10.1080/10618600.2024.2362232 2025

  16. [16]

    Informatica , author =

    From. Informatica , author =. 2011 , pages =. doi:10.15388/Informatica.2011.311 , language =

    work page doi:10.15388/informatica.2011.311 2011

  17. [17]

    Biometrical Journal , author =

    Multiset sparse redundancy analysis for high‐dimensional omics data , volume =. Biometrical Journal , author =. 2019 , pages =. doi:10.1002/bimj.201700248 , abstract =

    work page doi:10.1002/bimj.201700248 2019

  18. [18]

    Journal of Chemometrics , author =

    Multiblock redundancy analysis: interpretation tools and application in epidemiology , volume =. Journal of Chemometrics , author =. 2011 , pages =. doi:10.1002/cem.1392 , abstract =

    work page doi:10.1002/cem.1392 2011

  19. [19]

    Bioinformatics , author =

    Sparse redundancy analysis of high-dimensional genetic and genomic data , volume =. Bioinformatics , author =. 2017 , pages =. doi:10.1093/bioinformatics/btx374 , abstract =

    work page doi:10.1093/bioinformatics/btx374 2017

  20. [20]

    Advances in Data Analysis and Classification , author =

    View selection in multi-view stacking: choosing the meta-learner , volume =. Advances in Data Analysis and Classification , author =. 2025 , pages =. doi:10.1007/s11634-024-00587-5 , abstract =

    work page doi:10.1007/s11634-024-00587-5 2025

  21. [21]

    PloS One , author =

    A constrained singular value decomposition method that integrates sparsity and orthogonality , volume =. PloS One , author =. 2019 , keywords =. doi:10.1371/journal.pone.0211463 , abstract =

    work page doi:10.1371/journal.pone.0211463 2019

  22. [22]

    , title =

    Biclustering via. Biometrics , author =. 2010 , pages =. doi:10.1111/j.1541-0420.2010.01392.x , abstract =

    work page doi:10.1111/j.1541-0420.2010.01392.x 2010

  23. [23]

    Computational Statistics , author =

    Sparse reduced-rank regression for simultaneous rank and variable selection via manifold optimization , volume =. Computational Statistics , author =. 2023 , pages =. doi:10.1007/s00180-022-01216-5 , abstract =

    work page doi:10.1007/s00180-022-01216-5 2023

  24. [24]

    Friedman, Jerome and Hastie, Trevor and Tibshirani, Rob and Narasimhan, Balasubramanian and Tay, Kenneth and Simon, Noah and Yang, James , month = jun, year =. glmnet:. doi:10.32614/CRAN.package.glmnet , abstract =

    work page doi:10.32614/cran.package.glmnet

  25. [25]

    doi:10.32614/CRAN.package.RSpectra , abstract =

    Qiu, Yixuan and Mei, Jiali , month = feb, year =. doi:10.32614/CRAN.package.RSpectra , abstract =

    work page doi:10.32614/cran.package.rspectra

  26. [26]

    Nature neuroscience , author =

    The human connectome project's neuroimaging approach , volume =. Nature neuroscience , author =. 2016 , pages =

    2016

  27. [27]

    URL https://www.humanconnectome.org/storage/app/media/documentation/s1200/HCP1200-DenseConnectome+PTN+Appendix-July2017.pdf , author =

  28. [28]

    2016 , pages =

    Behavior Research Methods , author =. 2016 , pages =. doi:10.3758/s13428-015-0626-8 , language =

    work page doi:10.3758/s13428-015-0626-8 2016

  29. [29]

    Applied Mathematics Letters , author =

    An incremental randomized algorithm for singular value decomposition of streaming data matrices , volume =. Applied Mathematics Letters , author =. 2026 , pages =. doi:10.1016/j.aml.2025.109822 , language =

    work page doi:10.1016/j.aml.2025.109822 2026

  30. [30]

    Finding structure with randomness: Probabilistic algorithms for constructing approximate matrix decompositions

    Halko, Nathan and Martinsson, Per-Gunnar and Tropp, Joel A. , year =. Finding structure with randomness:. doi:10.48550/ARXIV.0909.4061 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.0909.4061

  31. [31]

    Perspectives on Psychological Science , author =

    Choosing. Perspectives on Psychological Science , author =. 2017 , pages =. doi:10.1177/1745691617693393 , abstract =

    work page doi:10.1177/1745691617693393 2017

  32. [32]

    Extending principal covariates regression for high-dimensional multi-block data , url =

    Park, Soogeun , year =. Extending principal covariates regression for high-dimensional multi-block data , url =

  33. [33]

    European Journal of Psychological Assessment , author =

    The. European Journal of Psychological Assessment , author =. 2015 , pages =. doi:10.1027/1015-5759/a000233 , abstract =

    work page doi:10.1027/1015-5759/a000233 2015

  34. [34]

    and Ettema, J.H.M

    Arrindell, W.A. and Ettema, J.H.M. and Derogatis, L.R. , year =

  35. [35]

    Computational Statistics & Data Analysis , author =

    Predictive performance of psychological tests: is it better to use items than subscales? , volume =. Computational Statistics & Data Analysis , author =. 2023 , pages =. doi:10.1016/j.csda.2023.107767 , language =

    work page doi:10.1016/j.csda.2023.107767 2023

  36. [36]

    European Journal of Psychological Assessment , author =

    The. European Journal of Psychological Assessment , author =. 2007 , pages =. doi:10.1027/1015-5759.23.3.141 , abstract =

    work page doi:10.1027/1015-5759.23.3.141 2007

  37. [37]

    Eigenvalue

    Zając, Katarzyna and Piersa, Jarosław , year =. Eigenvalue. Artificial Intelligence and Soft Computing , publisher =. doi:10.1007/978-3-642-38658-9_19 , urldate =

    work page doi:10.1007/978-3-642-38658-9_19

  38. [38]

    Principal component-guided sparse reduced-rank regression , url =

    Goto, Kanji and Yuki, Shintaro and Tanioka, Kensuke and Yadohisa, Hiroshi , month = jan, year =. Principal component-guided sparse reduced-rank regression , url =. doi:10.48550/arXiv.2601.07202 , abstract =

    work page doi:10.48550/arxiv.2601.07202

  39. [39]

    The Annals of Statistics , author =

    Estimation of high-dimensional low-rank matrices , volume =. The Annals of Statistics , author =. doi:10.1214/10-AOS860 , language =

    work page doi:10.1214/10-aos860

  40. [40]

    IEEE Transactions on Information Theory , author =

    Tight oracle inequalities for low-rank matrix recovery from a minimal number of noisy random measurements , volume =. IEEE Transactions on Information Theory , author =. 2011 , pages =. doi:10.1109/TIT.2011.2111771 , language =

    work page doi:10.1109/tit.2011.2111771 2011

  41. [41]

    The Annals of Statistics , author =

    Matrix estimation by universal singular value thresholding , volume =. The Annals of Statistics , author =. doi:10.1214/14-AOS1272 , language =

    work page doi:10.1214/14-aos1272

  42. [42]

    Constrained principal component analysis and related techniques , isbn =

    Takane, Yoshio , month = apr, year =. Constrained principal component analysis and related techniques , isbn =. doi:10.1201/b16020 , language =

    work page doi:10.1201/b16020

  43. [43]

    Statistical Analysis and Data Mining: The ASA Data Science Journal , author =

    Reduced rank ridge regression and its kernel extensions , volume =. Statistical Analysis and Data Mining: The ASA Data Science Journal , author =. 2011 , pages =. doi:10.1002/sam.10138 , abstract =

    work page doi:10.1002/sam.10138 2011

  44. [44]

    Biometrics , author =

    Hierarchical nuclear norm penalization for multi-view data integration , volume =. Biometrics , author =. 2023 , pages =. doi:10.1111/biom.13893 , abstract =

    work page doi:10.1111/biom.13893 2023

  45. [45]

    Advances in Data Analysis and Classification , author =

    Sparse constrained and unconstrained non-symmetric correspondence analysis , issn =. Advances in Data Analysis and Classification , author =. doi:10.1007/s11634-025-00646-5 , abstract =

    work page doi:10.1007/s11634-025-00646-5

  46. [46]

    European Respiratory Journal , author =

    Clinical, physiologic, imaging, and molecular responses to cannabis smoking: the canadian users of cannabis smoke (. European Respiratory Journal , author =. 2025 , pages =. doi:10.1183/13993003.01659-2025 , abstract =

    work page doi:10.1183/13993003.01659-2025 2025

  47. [47]

    Statistics & Probability Letters , author =

    Model selection and estimation in high dimensional regression models with group. Statistics & Probability Letters , author =. 2015 , pages =. doi:10.1016/j.spl.2015.04.017 , language =

    work page doi:10.1016/j.spl.2015.04.017 2015

  48. [48]

    Minimal spanning tree algorithm for gamma-ray source detection in sparse photon images: cluster parameters and selection strategies

    A cornucopia of maximum likelihood algorithms , issn =. American Statistician , author =. 2025 , pages =. doi:10.1080/00031305.2025.2526535 , language =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1080/00031305.2025.2526535 2025

  49. [49]

    Computerized Medical Imaging and Graphics , author =

    Predictive sparse modeling of. Computerized Medical Imaging and Graphics , author =. 2015 , pages =. doi:10.1016/j.compmedimag.2015.03.007 , language =

    work page doi:10.1016/j.compmedimag.2015.03.007 2015

  50. [50]

    Journal of the Acoustical Society of America , author =

    Don't force it!. Journal of the Acoustical Society of America , author =. 2022 , pages =. doi:10.1121/10.0015201 , abstract =

    work page doi:10.1121/10.0015201 2022

  51. [51]

    LWT--Food Science and Technology , author =

    Prediction of sensory texture of cooked potatoes using uniaxial compression, near infrared spectroscopy and low. LWT--Food Science and Technology , author =. 2000 , pages =. doi:10.1006/fstl.1999.0623 , language =

    work page doi:10.1006/fstl.1999.0623 2000

  52. [52]

    , year =

    Seber, George Arthur Frederick and Lee, Alan J. , year =. Linear regression analysis , isbn =

  53. [53]

    Reduced rank regression of functional data , url =

    Ebaid, Rasha , month = may, year =. Reduced rank regression of functional data , url =

  54. [54]

    Journal of Computational and Graphical Statistics , author =

    Multivariate. Journal of Computational and Graphical Statistics , author =. 2022 , pages =. doi:10.1080/10618600.2021.1960850 , language =

    work page doi:10.1080/10618600.2021.1960850 2022

  55. [55]

    2021 , pages =

    Behavior Research Methods , author =. 2021 , pages =. doi:10.3758/s13428-020-01508-y , language =

    work page doi:10.3758/s13428-020-01508-y 2021

  56. [56]

    Machine Learning: Science and Technology , author =

    Structure-property maps with kernel principal covariates regression , volume =. Machine Learning: Science and Technology , author =. 2020 , pages =. doi:10.1088/2632-2153/aba9ef , abstract =

    work page doi:10.1088/2632-2153/aba9ef 2020

  57. [57]

    Singular value decomposition (

    Abdi, Hervé , year =. Singular value decomposition (. Encyclopedia of. doi:10.4135/9781412952644.n413 , language =

    work page doi:10.4135/9781412952644.n413

  58. [58]

    Expositiones Mathematicae , author =

    von neumann’s trace inequality for hilbert–schmidt operators , volume =. Expositiones Mathematicae , author =. 2021 , pages =. doi:10.1016/j.exmath.2020.05.001 , language =

    work page doi:10.1016/j.exmath.2020.05.001 2021

  59. [59]

    Machine Learning , author =

    Variable selection for both outcomes and predictors: sparse multivariate principal covariates regression , volume =. Machine Learning , author =. 2024 , pages =. doi:10.1007/s10994-024-06520-3 , abstract =

    work page doi:10.1007/s10994-024-06520-3 2024

  60. [60]

    2nd ed., Springer, New York

    Hastie, Trevor and Tibshirani, Robert and Friedman, Jerome , year =. The elements of statistical learning , copyright =. doi:10.1007/978-0-387-84858-7 , language =

    work page doi:10.1007/978-0-387-84858-7

  61. [61]

    The effective rank: a measure of effective dimensionality , isbn =

    Roy, Olivier and Vetterli, Martin , year =. The effective rank: a measure of effective dimensionality , isbn =

  62. [62]

    and Omidvarnia, Amir and Küppers, Vincent and Patil, Kaustubh R

    Gell, Martin and Eickhoff, Simon B. and Omidvarnia, Amir and Küppers, Vincent and Patil, Kaustubh R. and Satterthwaite, Theodore D. and Müller, Veronika I. and Langner, Robert , month = feb, year =. The burden of reliability: how measurement noise limits brain-behaviour predictions , shorttitle =. doi:10.1101/2023.02.09.527898 , abstract =

    work page doi:10.1101/2023.02.09.527898 2023

  63. [63]

    Particle

    Sparse partial least squares regression for simultaneous dimension reduction and variable selection , volume =. Journal of the Royal Statistical Society Series B (Statistical Methodology) , author =. 2010 , pages =. doi:10.1111/j.1467-9868.2009.00723.x , abstract =

    work page doi:10.1111/j.1467-9868.2009.00723.x 2010

  64. [64]

    British Journal of Mathematical and Statistical Psychology , author =

    Simultaneous analysis of coupled data matrices subject to different amounts of noise , volume =. British Journal of Mathematical and Statistical Psychology , author =. 2011 , pages =. doi:10.1348/000711010X513263 , abstract =

    work page doi:10.1348/000711010x513263 2011

  65. [65]

    British Journal of Mathematical and Statistical Psychology , author =

    Selecting among three‐mode principal component models of different types and complexities: a numerical convex hull based method , volume =. British Journal of Mathematical and Statistical Psychology , author =. 2006 , pages =. doi:10.1348/000711005X64817 , abstract =

    work page doi:10.1348/000711005x64817 2006

  66. [66]

    Jinglong Zhao

    Ridge regression: biased estimation for nonorthogonal problems , volume =. Technometrics , author =. 1970 , pages =. doi:10.1080/00401706.1970.10488634 , language =

    work page doi:10.1080/00401706.1970.10488634 1970

  67. [67]

    Nature , author =

    Reproducible brain-wide association studies require thousands of individuals , volume =. Nature , author =. 2022 , pages =. doi:10.1038/s41586-022-04492-9 , abstract =

    work page doi:10.1038/s41586-022-04492-9 2022

  68. [68]

    Journal of the Royal Statistical Society Series B (Statistical Methodology) , author =

    Regularized matrix regression , volume =. Journal of the Royal Statistical Society Series B (Statistical Methodology) , author =. 2014 , pages =. doi:10.1111/rssb.12031 , abstract =

    work page doi:10.1111/rssb.12031 2014

  69. [69]

    Journal of the Royal Statistical Society Series B: Statistical Methodology , volume =

    Regularization and variable selection via the elastic net , volume =. Journal of the Royal Statistical Society Series B (Statistical Methodology) , author =. 2005 , pages =. doi:10.1111/j.1467-9868.2005.00503.x , abstract =

    work page doi:10.1111/j.1467-9868.2005.00503.x 2005

  70. [70]

    Journal of the Royal Statistical Society: Series B (Methodological)58(1), 267–288 (1996) https://doi

    Regression shrinkage and selection via the lasso , volume =. Journal of the Royal Statistical Society Series B (Statistical Methodology) , author =. 1996 , pages =. doi:10.1111/j.2517-6161.1996.tb02080.x , abstract =

    work page doi:10.1111/j.2517-6161.1996.tb02080.x 1996

  71. [71]

    Psychometrika , author =

    Redundancy analysis an alternative for canonical correlation analysis , volume =. Psychometrika , author =. 1977 , pages =. doi:10.1007/BF02294050 , abstract =

    work page doi:10.1007/bf02294050 1977

  72. [72]

    doi:https://doi.org/10.1016/0047-259X(75)90042-1 , issn =

    Reduced-rank regression for the multivariate linear model , volume =. Journal of Multivariate Analysis , author =. 1975 , pages =. doi:10.1016/0047-259X(75)90042-1 , language =

    work page doi:10.1016/0047-259x(75)90042-1 1975

  73. [73]

    British Journal of Mathematical and Statistical Psychology , author =

    Reduced rank regression for mixed predictor and response variables , issn =. British Journal of Mathematical and Statistical Psychology , author =. 2025 , pages =. doi:10.1111/bmsp.70004 , abstract =

    work page doi:10.1111/bmsp.70004 2025

  74. [74]

    Applied Statistics , author =

    Procedures for reduced-rank regression , volume =. Applied Statistics , author =. 1982 , pages =. doi:10.2307/2347998 , language =

    work page doi:10.2307/2347998 1982

  75. [75]

    Chemometrics and Intelligent Laboratory Systems , author =

    Principal covariates regression: part. Chemometrics and Intelligent Laboratory Systems , author =. 1992 , pages =. doi:10.1016/0169-7439(92)80100-I , language =

    work page doi:10.1016/0169-7439(92)80100-i 1992

  76. [76]

    Nature Human Behaviour , author =

    Power and reproducibility in the external validation of brain-phenotype predictions , volume =. Nature Human Behaviour , author =. 2024 , pages =. doi:10.1038/s41562-024-01931-7 , language =

    work page doi:10.1038/s41562-024-01931-7 2024

  77. [77]

    doi:10.18637/jss.v065.i08 , language =

    Journal of Statistical Software , author =. doi:10.18637/jss.v065.i08 , language =

    work page doi:10.18637/jss.v065.i08

  78. [78]

    Annals of Statistics , author =

    Optimal selection of reduced rank estimators of high-dimensional matrices , volume =. Annals of Statistics , author =. doi:10.1214/11-AOS876 , language =

    work page doi:10.1214/11-aos876

  79. [79]

    Mathematische Nachrichten , author =

    On the trace of matrix products , volume =. Mathematische Nachrichten , author =. 1959 , pages =. doi:10.1002/mana.19590200306 , language =

    work page doi:10.1002/mana.19590200306 1959

  80. [80]

    Chemometrics and Intelligent Laboratory Systems , author =

    On the selection of the weighting parameter value in principal covariates regression , volume =. Chemometrics and Intelligent Laboratory Systems , author =. 2013 , pages =. doi:10.1016/j.chemolab.2013.02.005 , language =

    work page doi:10.1016/j.chemolab.2013.02.005 2013

Showing first 80 references.