Pith. sign in

REVIEW 3 major objections 6 minor 47 references

Generalized Neural Distributional Regression

T0 review · 3 major / 6 minor · reviewed 2026-08-02 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read By freezing a deep network's hidden layers as a fixed basis expansion, the terminal layer becomes a classical MLE with recoverable Fisher information — observation-specific confidence bands follow for any distributional parameter.

desk verdict A genuinely useful two-step MLE framework with a real conditional-inference caveat: the reported intervals ignore first-step feature-estimator uncertainty, and the test-set model selection weakens the empirical claims. read the letter →

arxiv 2607.14122 v1 pith:DDIBRT2A submitted 2026-06-16 stat.ML cs.LGmath.STstat.APstat.COstat.MEstat.TH

Natan Hilario da Silva , Vicente Garibay Cancho , Adriano Kamimura Suzuki This is my paper · ORCID
classification stat.MLcs.LGmath.STstat.APstat.COstat.MEstat.TH MSC 62F1268T07
keywords neuraldistributionalregressionuncertaintyquantificationFisherinformationmatrixtwo-stepestimationDeltamethodasymptoticnormalitysurvivalanalysisdeeplearning
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper proposes Generalized Neural Distributional Regression (GNDR), a way to embed deep neural networks inside classical parametric distributions, modeling each distributional parameter as a link-transformed network output. Its central claim is that the non-identifiability that normally blocks statistical inference in deep networks can be sidestepped by a two-step scheme: first train the whole network, then freeze all hidden layers and re-fit only the final output layer with full-batch maximum likelihood. In the second step the network is mathematically a vector generalized linear model over a learned feature map, so the estimator is asymptotically normal and the observed Hessian is a valid covariance matrix. This makes it possible to attach valid, per-observation confidence intervals to parameters, quantiles, and survival curves even when the input is an image or a high-dimensional gene-expression vector. Empirical results show the framework matching or improving on classical additive distributional-regression baselines in calibration while processing unstructured data.

What carries the argument

The load-bearing device is the two-step estimation split ω = {ω_deep, ω_last}. Step 1 optimizes the full network by stochastic gradient methods; step 2 permanently freezes ω̂_deep, so the final hidden layer becomes a deterministic learned basis expansion v(x) = F_deep(x; ω̂_deep), and the network output collapses to the linear form F(x; ω) = W_lastᵀ v(x) + b_last. With the structured effects ξ, the model is then exactly a vector generalized linear model — a linear model for several distribution parameters at once — over the identifiable parameter vector θ_last = (ξ, vec(W_last), b_last), whose design matrix [z, v(x)] is full rank thanks to the disjoint-covariate condition. The observed Fishe

What would settle it

Simulate data from a GNDR model whose true parameters are known, with a wide terminal bottleneck so the learned features carry nontrivial estimation error. Fit the two-step procedure across many Monte Carlo replicates and compare the empirical standard deviation of θ̂_last with the mean inverse observed Hessian: if the ratio deviates from one, or if Delta-method interval coverage falls below nominal while a full two-step bootstrap holds nominal coverage, then Theorem 1's conditional characterization is incomplete.

Watch

Extended reading notes

Core claim

Conditional on the frozen latent feature vectors v(x_i) produced by the upstream network, the GNDR estimator θ̂_last is asymptotically normal: √n(θ̂_last − θ_last⁰) → N_q(0, Σ_last) with Σ_last = I⁻¹(θ_last⁰) (Theorem 1). Because the second training step drives the restricted log-likelihood to a strict root of the score equation over a fixed, finite-dimensional parameter space, standard regularity conditions for vector generalized linear models apply, and the inverse observed Hessian (Eq. 8) is a machine-precision-accurate covariance estimate. From this anchor, the paper chains three Delta-method applications — through the data-dependent projection P(x), the predictor mapping h, and the inve

Load-bearing premise

The entire inferential output treats the deep features from the first training step as fixed and known, so all confidence intervals are conditional on them — the paper's own facial-image experiment finds 89.35% out-of-sample coverage for nominal 95%, which is the size of the leak when that premise is strained.

Editorial extensions

If this is right

  • For any base distribution with a smooth, twice-differentiable log-likelihood, the framework yields observation-specific confidence bands for parameters, quantiles, survival curves, and tolerance intervals, all computed from one analytical covariance matrix (Eqs. 8, 13–16).
  • GNDR reduces to classical additive distributional regression when the neural component is zeroed out, while the neural version can capture interaction surfaces that independent penalized splines miss; the paper's hospital-stay application shows the neural interaction model giving the best out-of-sample calibration.
  • The theory changes model-selection practice: in-sample AIC/BIC and global deviance are declared invalid for comparing deep distributional architectures, replaced by out-of-sample Anderson-Darling statistics on normalized quantile residuals — with the caveat, in censored survival data, that residual diagnostics be paired with visual checks of predicted survival curves.
  • Architecture design follows from the theory: the terminal hidden layer should be an aggressive bottleneck (width M ≤ 4 in the applications), because the second-step parameter count — b + M p_F + p_F — is what determines statistical degrees of freedom and variance inflation.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The intervals in the paper are conditional intervals: they treat the learned features from the first step as fixed, so the honest test of the method is to compare their frequentist coverage with a full two-step bootstrap; the paper's own image experiment (89.35% out-of-sample coverage for nominal 95%) suggests the conditioning gap is material at that scale.
  • The width of the terminal bottleneck plays a dual role — representation capacity in step one, statistical degrees of freedom in step two — which points toward a sieve-style asymptotic theory in which M grows slowly with n; that formalization is implicit in the paper but not developed.
  • The freeze-and-refit device should transfer to other inference paradigms: Bayesian terminal-layer inference (where the observed Hessian doubles as a Laplace posterior covariance), neural random effects for clustered data, and conformal wraps around the Delta-method intervals as a finite-sample correction.
  • A direct calibration experiment is easy to run: simulate from a known GNDR model, refit the full two-step procedure many times, and check whether the inverse observed Hessian tracks the Monte Carlo variance of the terminal-layer estimator — if it does not, the gap identifies exactly where the conditional-normality premise leaks.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The paper introduces Generalized Neural Distributional Regression (GNDR), a two-step estimation framework for parametric distributional regression with deep neural networks. In the first step, the full network (including the terminal layer) is trained by stochastic optimization with early stopping. In the second step, all internal layers are frozen and only the terminal dense layer and classical structured effects are re-estimated by full-data maximum likelihood in a reduced Vector Generalized Linear Model (VGLM). Theorem 1 claims conditional asymptotic normality of this terminal-layer estimator, with covariance given by the inverse Fisher information; the Delta method is then used to build observation-specific confidence bands for distributional parameters, quantiles, and survival curves. The framework is demonstrated on four applications: Dutch boys' BMI with a Box-Cox t distribution, hospital stay counts with a Beta-binomial distribution, TCGA-BRCA survival with various parametric and mixture-cure models, and age prediction from UTKFace images with a zero-truncated normal distribution. The paper also presents the open-source package 'thetaflow'.

Significance. If the conditional-normality claim could be upgraded to a valid unconditional two-step inference statement, GNDR would be a practically valuable bridge between deep representation learning and classical parametric inference, enabling uncertainty quantification on unstructured/high-dimensional data. The manuscript has several strengths: the identifiability discussion is explicit, the reduction to a VGLM is transparent, the Hessian is computed analytically via automatic differentiation, and the authors provide reproducible code and a package. The empirical scope is broad, covering tabular, genomic, and image data. However, the central UQ promise is not yet justified: the asymptotic covariance in Eq. (8) ignores the randomness of the first-step feature extractor, and the model-selection protocol in Section 3.3 makes the reported test-set calibration partly self-referential. These issues are load-bearing for the paper's main claims.

major comments (3)
  1. [§3.1–3.2, Theorem 1, Eq. (8)] Theorem 1 conditions on v(x)=F_deep(x; \hat\omega_deep), but \hat\omega_deep is estimated from the same training responses and is correlated with the second-step likelihood. The covariance in Eq. (8) is the conditional Fisher information given the realized features, not the total variance of \hat\theta_last or of the Delta-method bands in Eqs. (11)–(16). Standard two-step inference (e.g., Murphy–Topel or Newey–McFadden corrections) is not applied. The empirical test-set coverage of 89.35% in §4.4 is exactly the direction expected from this omission. The manuscript should either provide a corrected covariance that propagates first-step uncertainty, or explicitly reframe the UQ as conditional on the estimated features and temper the coverage claims.
  2. [§3.3, §4.2, §4.3] The model-selection rule in §3.3 selects the model minimizing the Anderson–Darling statistic on the test set. The applications then report the test-set AD/calibration of the selected model as evidence of 'superior distributional calibration' (e.g., Table 2 and the claim that model (d) 'emerges as the superior architecture'). This is self-referential: the test set is used for selection, so the reported test performance is optimistically biased. A separate validation set for selection, or nested cross-validation, is needed for honest out-of-sample evaluation.
  3. [Appendix A] The proof of Theorem 1 verifies the VGLM regularity conditions by assertion rather than by argument. In particular, Regularity Condition II requires the concatenated design matrix [z, v(x)] to have full column rank, but v(x)=F_deep(x; \hat\omega_deep) is itself random and estimated, so the fixed-design VGLM conditions of Yee (2015) are not automatically satisfied. The proof also does not clarify the definition of the 'true' θ_last^0 when the feature map is estimated. A rigorous proof requires either an explicit asymptotic framework for the two-step estimator (treating \hat\omega_deep as a nuisance function) or a sample-splitting scheme that makes the feature extractor independent of the second-step likelihood.
minor comments (6)
  1. [Abstract] Typo: 'packagethetaflow' should read 'package thetaflow'.
  2. [§2.2] Heading typo: 'identifiabiliy' should be 'identifiability'.
  3. [Eq. (16)] The notation in (16) writes the limiting distribution as N_p, but f maps R^p to R^q; the dimension should be q or the notation should be clarified.
  4. [§4.1] The Box-Cox t log-likelihood uses f_T and F_T without explicitly stating that T denotes a Student-t distribution with τ_i degrees of freedom; please define these symbols.
  5. [Table 3] The 'logit(λ/40)' link for the Cure Mixture Weibull scale imposes an identifiability bound; it would help to state explicitly that this is a modeling assumption and how sensitive results are to the choice of 40.
  6. [§3.3] The sentence about AIC/BIC being invalid 'except they are comparing the exact same distribution, as it happens in cross validation' is confusing; please clarify the intended scope of the advice.

Circularity Check

1 steps flagged · score 3.0 of 10

Central UQ theory is not circular, but the reported 'superior distributional calibration' is partly self-confirming: the same test-set Anderson–Darling statistic is used both to select the best model and as the evidence that it is best.

  1. fitted input called prediction [Model selection and evaluation: Section 3.3; Section 4.2 (Table 2); Section 4.3 (Table 4)]
    "Therefore, when evaluating a set of candidate GNDR architectures or base distributions on the test set, the model minimizing the AD statistic is preferred. ... Based on the out-of-sample evaluation established in Section 3.3, model (d) emerges as the superior architecture, minimizing the AD statistic on the test set. ... Table 4 presents the AD test statistic for each candidate architecture, identifying the Log-logistic distribution as the optimal fit."

    The Anderson–Darling statistic on the held-out test set is both the model-selection criterion (Section 3.3: 'the model minimizing the AD statistic is preferred') and the subsequently reported evidence of superiority (Tables 2 and 4, and the surrounding text). Because the chosen model is selected by minimizing exactly this test-set quantity, its top ranking on that same quantity is guaranteed by the selection rule rather than independently confirmed. The test set thus functions as part of the selection procedure, so the empirical claim that the selected GNDR model has superior distributional calibration is partly forced by construction. This does not make Theorem 1 circular, but it weakens the empirical demonstration.

full rationale

The core mathematical derivation is not circular. Theorem 1 explicitly conditions on the frozen latent features v(x), and the proof verifies standard VGLM regularity conditions for the conditional second-step model. Within that conditional model, θ_last is an identifiable parameter, the likelihood is well-defined, and the Fisher information and Delta-method covariance calculations follow from standard asymptotic theory. The fact that v(x) is itself learned from the same responses is a real limitation—the reported intervals are conditional on the fitted feature extractor and omit first-step estimation error, which is consistent with the observed out-of-sample coverage of 89.35%—but it is an incompleteness/statistical weakness, not a circular reduction. The paper also cites prior work for LLLA, neural distributional regression, and VGLMs without relying on a self-citation chain. The only concrete circularity-like step is the empirical model-selection/evaluation loop: the test-set AD statistic is used both to choose the best architecture and as the headline evidence of superior calibration. That step is statistically self-referential and inflates the apparent evidence, but it does not affect the central asymptotic theorem. Accordingly, the overall circularity score is modest.

Assumptions & free parameters 7 free parameters · 7 assumptions · 0 invented entities

The derivation's main input is the assumption that the frozen deep features v(x) form a fixed basis; this is what makes θ_last an identifiable finite-dimensional VGLM parameter. Everything else (smoothness, rank, identifiability of D) is standard regularity. The empirical sections add ad hoc constraints (λ≤40) and test-set model selection, which are not carried through the covariance.

free parameters (7)
  • Terminal hidden-layer width M = 1 (hospital), 2 (Dutch boys/survival), 64 (UTKFace)
    Chosen by hand per application; determines q, FIM conditioning, and overfitting risk. §3.1 recommends M≤4, but UTKFace uses M=64.
  • Dropout rate = 0.1–0.45 (TCGA), 0.3 (UTKFace)
    Selected via 5-fold CV on training data for survival; fixed by heuristic for faces. Affects estimates and coverage but is not included in the covariance.
  • Elastic Net L1/L2 penalties = L1 in {0, 0.0001, 0.0005, 0.1}; L2 in {0.0001, 0.005, 0.01}
    Per-model cross-validated values in Table 3; hyperparameters fitted to training data, not propagated into the FIM.
  • Cure-mixture Weibull scale upper bound = 40
    Ad hoc bound imposed through logit(λ/40) in Table 3; prevents degenerate λ→∞ and p→0 but is a modeling constraint, not derived from data.
  • First-stage epoch/early-stopping schedule = Not specified
    Validation-based early stopping and learning-rate decay control convergence; no reproducible schedule or seed is given, and this affects the frozen features.
  • First hidden-layer width in survival network = 32
    Chosen by 'empirical diagnostics' (§4.3) rather than a formal criterion; changes the learned basis and downstream inference.
  • Network topology/activations = e.g., 24-2 GELU, 24-1 GELU, 32-2 tanh/softplus, MobileNetV2 128/64 GELU
    Architectural choices define the basis expansion and are tuned per application; they are not treated as statistical parameters.
assumptions (7)
  • standard math MLE regularity conditions for VGLMs hold for the chosen base distributions and links.
    Appendix A verifies these by assertion (fixed dimension, identifiability, common support, open parameter space, smoothness/information). No formal check is given, especially for cure-mixture models.
  • domain assumption After step 1, v(x) = F_deep(x; ω̂_deep) can be treated as a fixed, known basis expansion.
    §3.1 freezes all internal layers; Theorem 1 is conditional on v(x). All covariance statements in the applications treat this as unconditional, which is the paper's chief unacknowledged assumption.
  • domain assumption The augmented design matrix [Z, v(X)] has full column rank.
    §3.1 states this must hold to avoid collinearity, but no rank diagnostics are reported for any application.
  • domain assumption Neural and classical features feeding the same distributional parameter are disjoint.
    §2.2 imposes disjointness for identifiability; in the shared-implementation models the global input x may contain all covariates, so this is only guaranteed for the reported z choices.
  • standard math The base distribution D is identifiable and has support independent of parameters.
    Definition 1 and Regularity Condition III; for the mixture cure Weibull an extra bound is required to make the parametrization identifiable.
  • ad hoc to paper The bound λ ∈ (0,40) fixes mixture-cure Weibull identifiability.
    Table 3 maps logit(λ/40); the upper bound 40 is a biological plausibility constraint set by hand, not estimated or derived from the data.
  • domain assumption Selecting the best model by test-set AD statistic leaves an unbiased performance evaluation.
    §3.3 recommends AD-based model selection, but Tables 2 and 4 then report test AD for the selected model, so selection and evaluation share the same data.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Generalized Neural Distributional Regression." pith.science (2026). https://pith.science/paper/DDIBRT2A

@misc{pith2026260714122,
  author       = {Pith},
  title        = {Pith review of: Generalized Neural Distributional Regression},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DDIBRT2A}},
  note         = {Machine review of arXiv:2607.14122}
}
read the original abstract

We introduce the Generalized Neural Distributional Regression (GNDR) framework, which seamlessly embeds deep neural networks into the parameter space of classical probability distributions. To reconcile the inherent non-identifiability of deep architectures with maximum likelihood theory, we propose a two-step semi-parametric estimation procedure. By isolating the terminal prediction heads and treating the upstream network as a fixed, non-linear basis expansion, GNDR enables the extraction of analytical Fisher Information matrices. This facilitates rigorous uncertainty quantification, generating observation-specific confidence bands and tolerance intervals via the multivariate Delta method. We demonstrate the framework's versatility and superior distributional calibration across diverse data modalities, including overdispersed clinical counts, right-censored transcriptomic survival profiles under a mixture cure framework, and zero-truncated age distributions derived directly from unstructured facial images. The methodology is natively implemented in the open-source Python package \textit{thetaflow}.

Figures

Figures reproduced from arXiv: 2607.14122 by the authors.

Figure 1
Figure 1. Four natural regression models obtained from a [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. BCT parameters against age, comparing the classical penalized spline GAMLSS [PITH_FULL_IMAGE:figures/full_fig_p024_2.png] view at source ↗
Figure 3
Figure 3. Q-Q plots and worm plots for both the GNDR and GAMLSS models for the [PITH_FULL_IMAGE:figures/full_fig_p025_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Surface plot illustrating the learned non-linear neural interaction effect ( [PITH_FULL_IMAGE:figures/full_fig_p028_4.png]
Figure 5
Figure 5. Figure 5: Wormplots for train and test sets for both models (b) and (d) in the Hospital [PITH_FULL_IMAGE:figures/full_fig_p029_5.png]
Figure 6
Figure 6. Figure 6: Predicted survival curves, alongside their average curve juxtaposed to the Kaplan [PITH_FULL_IMAGE:figures/full_fig_p035_6.png]
Figure 7
Figure 7. Figure 7: Nelson-Aalen cumulative hazard estimator of the Cox-Snell residuals for the op [PITH_FULL_IMAGE:figures/full_fig_p036_7.png]
Figure 8
Figure 8. Figure 8: Worm plots of the randomized normalized quantile residuals for the Log-logistic [PITH_FULL_IMAGE:figures/full_fig_p037_8.png]
Figure 9
Figure 9. Figure 9: Patient-specific predicted survival curves for three distinctly profiled individuals [PITH_FULL_IMAGE:figures/full_fig_p038_9.png]
Figure 10
Figure 10. Figure 10: Out-of-sample age prediction for six diverse subjects from the UTKFace test [PITH_FULL_IMAGE:figures/full_fig_p042_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

47 extracted references · 3 linked inside Pith

  1. [1]

    2014 Fourth international conference on advanced computing & communication technologies , pages=

    A detailed review of feature extraction in image processing systems , author=. 2014 Fourth international conference on advanced computing & communication technologies , pages=. 2014 , organization=

    2014

  2. [2]

    Nature Reviews Methods Primers , volume=

    Principal component analysis , author=. Nature Reviews Methods Primers , volume=. 2022 , publisher=

    2022

  3. [3]

    Machine learning for data science handbook: data mining and knowledge discovery handbook , pages=

    Autoencoders , author=. Machine learning for data science handbook: data mining and knowledge discovery handbook , pages=. 2023 , publisher=

    2023

  4. [4]

    Information fusion , volume=

    A review of uncertainty quantification in deep learning: Techniques, applications and challenges , author=. Information fusion , volume=. 2021 , publisher=

    2021

  5. [5]

    ACM Computing Surveys , volume=

    A survey on uncertainty quantification methods for deep learning , author=. ACM Computing Surveys , volume=. 2026 , publisher=

    2026

  6. [6]

    , author=

    A tutorial on conformal prediction. , author=. Journal of machine learning research , volume=

  7. [7]

    2012 , publisher=

    Bayesian learning for neural networks , author=. 2012 , publisher=

    2012

  8. [8]

    Artificial Intelligence Review , volume=

    Deep learning for survival analysis: a review , author=. Artificial Intelligence Review , volume=. 2024 , publisher=

    2024

Show all 47 references
  1. [9]

    ACM Computing Surveys (CSUR) , volume=

    Machine learning for survival analysis: A survey , author=. ACM Computing Surveys (CSUR) , volume=. 2019 , publisher=

    2019

  2. [10]

    International Journal of Machine Learning and Cybernetics , volume=

    Deep learning for time series forecasting: a survey , author=. International Journal of Machine Learning and Cybernetics , volume=. 2025 , publisher=

    2025

  3. [11]

    Research , volume=

    Causal inference meets deep learning: A comprehensive survey , author=. Research , volume=. 2024 , publisher=

    2024

  4. [12]

    IEEE transactions on knowledge and data engineering , volume=

    Deep learning for spatio-temporal data mining: A survey , author=. IEEE transactions on knowledge and data engineering , volume=. 2020 , publisher=

    2020

  5. [13]

    Journal of the Royal Statistical Society Series C: Applied Statistics , volume=

    Generalized additive models for location, scale and shape , author=. Journal of the Royal Statistical Society Series C: Applied Statistics , volume=. 2005 , publisher=

    2005

  6. [14]

    Computational Statistics & Data Analysis , volume=

    Deep distribution regression , author=. Computational Statistics & Data Analysis , volume=. 2021 , publisher=

    2021

  7. [15]

    arXiv preprint arXiv:2010.06889 , year=

    Neural mixture distributional regression , author=. arXiv preprint arXiv:2010.06889 , year=

    2010 arXiv

  8. [16]

    2000 , publisher=

    Methods of information geometry , author=. 2000 , publisher=

    2000

  9. [17]

    Constructive Approximation , volume=

    Neural network identifiability for a family of sigmoidal nonlinearities , author=. Constructive Approximation , volume=. 2022 , publisher=

    2022

  10. [18]

    Neurocomputing , volume=

    Structural identifiability of generalized constraint neural network models for nonlinear regression , author=. Neurocomputing , volume=. 2008 , publisher=

    2008

  11. [19]

    NIPS workshop on bayesian deep learning , volume=

    Improving the identifiability of neural networks for Bayesian inference , author=. NIPS workshop on bayesian deep learning , volume=

  12. [20]

    Machine learning , volume=

    Parameter identifiability of a deep feedforward ReLU neural network , author=. Machine learning , volume=. 2023 , publisher=

    2023

  13. [21]

    arXiv preprint arXiv:1412.6980 , year=

    Adam: A method for stochastic optimization , author=. arXiv preprint arXiv:1412.6980 , year=

    arXiv

  14. [22]

    arXiv preprint arXiv:1711.05101 , volume=

    Fixing weight decay regularization in adam , author=. arXiv preprint arXiv:1711.05101 , volume=

    arXiv

  15. [23]

    2006 , publisher=

    Pattern recognition and machine learning , author=. 2006 , publisher=

    2006

  16. [24]

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

    Vector generalized additive models , author=. Journal of the Royal Statistical Society Series B: Statistical Methodology , volume=. 1996 , publisher=

    1996

  17. [25]

    Advances in neural information processing systems , volume=

    Laplace redux-effortless bayesian deep learning , author=. Advances in neural information processing systems , volume=

  18. [26]

    1983 , organization=

    Information measures and model selection , author=. 1983 , organization=

    1983

  19. [27]

    Journal of Computational and graphical statistics , volume=

    Randomized quantile residuals , author=. Journal of Computational and graphical statistics , volume=. 1996 , publisher=

    1996

  20. [28]

    Statistics in medicine , volume=

    Worm plot: a simple diagnostic device for modelling growth reference curves , author=. Statistics in medicine , volume=. 2001 , publisher=

    2001

  21. [29]

    goodness of fit

    Asymptotic theory of certain" goodness of fit" criteria based on stochastic processes , author=. The annals of mathematical statistics , pages=. 1952 , publisher=

    1952

  22. [30]

    Statistics in medicine , volume=

    On the C-statistics for evaluating overall adequacy of risk prediction procedures with censored survival data , author=. Statistics in medicine , volume=. 2011 , publisher=

    2011

  23. [31]

    Statistical Modelling , volume=

    Using the Box-Cox t distribution in GAMLSS to model skewness and kurtosis , author=. Statistical Modelling , volume=. 2006 , publisher=

    2006

  24. [32]

    Journal of the Royal Statistical Society Series C: Applied Statistics , volume=

    Use of the beta--binomial distribution to model the effect of policy changes on appropriateness of hospital stays , author=. Journal of the Royal Statistical Society Series C: Applied Statistics , volume=. 1996 , publisher=

    1996

  25. [33]

    PLoS computational biology , volume=

    Cox-nnet: an artificial neural network method for prognosis prediction of high-throughput omics data , author=. PLoS computational biology , volume=. 2018 , publisher=

    2018

  26. [34]

    Journal of the royal statistical society: Series B (methodological) , volume=

    Regression models and life-tables , author=. Journal of the royal statistical society: Series B (methodological) , volume=. 1972 , publisher=

    1972

  27. [35]

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

    Regularization and variable selection via the elastic net , author=. Journal of the Royal Statistical Society Series B: Statistical Methodology , volume=. 2005 , publisher=

    2005

  28. [36]

    Journal of the American Statistical Association , volume=

    Survival curve for cancer patients following treatment , author=. Journal of the American Statistical Association , volume=. 1952 , publisher=

    1952

  29. [37]

    Statistics in medicine , volume=

    Model diagnostics for censored regression via randomized survival probabilities , author=. Statistics in medicine , volume=. 2021 , publisher=

    2021

  30. [38]

    Statistics in medicine , volume=

    Assessment and comparison of prognostic classification schemes for survival data , author=. Statistics in medicine , volume=. 1999 , publisher=

    1999

  31. [39]

    Jama , volume=

    Evaluating the yield of medical tests , author=. Jama , volume=. 1982 , publisher=

    1982

  32. [40]

    Nucleic Acids Research , volume=

    KEGG: Kyoto Encyclopedia of Genes and Genomes , author=. Nucleic Acids Research , volume=. 2000 , publisher=

    2000

  33. [41]

    Frontiers in Genetics , volume=

    Identification and validation of a seven-gene metastasis-associated prognostic model in breast cancer , author=. Frontiers in Genetics , volume=. 2026 , publisher=

    2026

  34. [42]

    MedComm , volume=

    Breast Cancer: Molecular Pathogenesis and Targeted Therapy , author=. MedComm , volume=. 2025 , publisher=

    2025

  35. [43]

    International Journal of Medical Science and Clinical Research Studies , volume=

    Clinical Applications of Multi-Omics Integration in Disease Diagnosis , author=. International Journal of Medical Science and Clinical Research Studies , volume=. 2026 , doi=

    2026

  36. [44]

    Biomolecules , volume=

    Challenges and Opportunities in Multi-Omics Data Acquisition and Analysis: Toward Integrative Solutions , author=. Biomolecules , volume=. 2026 , publisher=

    2026

  37. [45]

    Proceedings of the IEEE International Conference on Computer Vision , pages=

    Age progression/regression by conditional adversarial autoencoder , author=. Proceedings of the IEEE International Conference on Computer Vision , pages=

  38. [46]

    Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

    Mobilenetv2: Inverted residuals and linear bottlenecks , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

  39. [47]

    2015 , publisher=

    Vector generalized linear and additive models: with an implementation in R , author=. 2015 , publisher=

    2015

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.