Pith. sign in

REVIEW 4 major objections 5 minor 37 references

Population-aware Hierarchical Bayesian Domain Adaptation via Multiple-component Invariant Learning

T0 review · 4 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read A hierarchical Bayesian model improves influenza prediction when target data come from a new environment and a different population.

desk verdict A useful hierarchical transfer method with a shaky causal story; deserves review after fixes. read the letter →

arxiv 1908.09222 v5 pith:3WDRZPV2 submitted 2019-08-24 stat.ML cs.LG

Vishwali Mhasawade , Nabeel Abdur Rehman , Rumi Chunara This is my paper · ORCID
classification stat.MLcs.LG
keywords domainadaptationhierarchicalBayesianmodelinvariantlearningobservationaltransportselectiondiagraminfluenzapredictionrepresentationbiaspopulationsubgroups
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

Flu and similar health predictions often fail when a model built in one environment is applied to data gathered elsewhere: symptom reporting differs across collection modes, and demographic subgroups are represented unevenly. This paper argues that both problems can be handled in one model by treating the demographic conditional distribution P(Y|D) as invariant across environments and the symptom–infection mapping P*(Y|X,D) as environment-specific, a split justified by an explicit causal selection diagram. The proposed hierarchical Bayesian model (Hier+pop) learns shared age- and gender-level parameters at the top of the hierarchy and dataset-specific parameters at the bottom, letting each target subgroup borrow what it needs. On four influenza datasets, with only 20% of the target data labeled, it outperforms all baselines in every target environment, with AUC gains such as 0.744 versus 0.645 on Goviral and 0.754 versus 0.546 on Fluwatch. If the causal assumptions hold, the same structure could reduce the labeling burden for many observational health prediction tasks.

What carries the argument

The central object is the selection diagram (a causal graph augmented with selection variables) plus a multi-parent hierarchical Bayesian model built to match it. In the diagram, selection variables S* mark features whose measurement changes across environments and S̃ marks demographic selection bias; from the graph the authors read that infection given demographics, P(Y|D), is invariant, while the symptom-to-infection mapping is not. The hierarchical model realizes this by placing a shared root parameter θ_pop above population parameters θ_a, θ_g, above environment parameters θ_c, above dataset parameters θ_l, and fitting all levels jointly with a maximum-a-posteriori objective in which an L2 divergence penalizes a child parameter for straying from its parents. A separate per-subgroup regression learns positive weights γ combining dataset, age, and gender parameters for each demographic subgroup, and Theorem 1 gives the licensing conditions (subgroup informativeness δ and prevalence gap) that decide whether the local θ_l or the invariant θ_d should dominate.

What would settle it

A decisive test would be a target environment deliberately chosen so that P(Y|D) differs from the pooled sources—for instance, an age group with opposite infection risk in the target, or a setting where laboratory confirmation is only available for severe cases. If Hier+pop, trained on source data plus 20% labeled target, does not outperform target-only training on that environment, the population-invariant component is not doing the claimed work. Equivalently, a simulation sampling from a graph with an unobserved confounder U affecting both selection and infection, violating S ⊥ Y | D, should reproduce the failure.

Watch

Extended reading notes

Core claim

The paper's central claim is that, for symptom-based infection prediction, demographic attributes define a transferable invariant and symptom-reporting behavior defines a non-transferable component. Using a selection diagram with two types of selection variables—one marking changes in how symptoms are assigned (S*→X) and one marking demographic selection bias (D→S̃)—the authors derive that S ⊥ Y | D but S ⊥̸ Y | X, so P(Y|D) can be carried across environments while P*(Y|X,D) must be re-learned locally. The model instantiates this split as a hierarchy whose population parameters (by age group and gender) sit above environment and dataset parameters, with a divergence penalty pulling children toward parents; per-subgroup weights decide how much each target subgroup relies on invariant versus local information. Experiments over four observational influenza datasets show large AUC gains when the target has only 20% labeled data, and subgroup results show gains even for underrepresented groups, with the paper's Theorem 1 giving conditions under which the model should fall back on dataset-specific parameters.

Load-bearing premise

The load-bearing premise is that the true process that created the data matches the paper's selection diagram—specifically that, once age and gender are known, being selected into a dataset has no effect on infection status, so the age- and gender-stratified infection rates can be carried from one environment to another; if unmeasured causes (such as differing lab-testing practices or exposure patterns) shift those rates, the invariant component transfers the wrong information.

Editorial extensions

If this is right

  • With only 20% labeled target data, Hier+pop achieves consistently higher AUC than target-only training, logistic regression, and FEDA-style feature-augmentation baselines on all four datasets, so practitioners can label less.
  • Population-invariant parameters improve prediction for demographic subgroups that are underrepresented in a target dataset, mitigating a form of aggregation bias.
  • The learned per-subgroup weights reveal when local environment data is needed: if the subgroup's symptoms are more informative than the pooled population's, or if its prevalence diverges sharply, the model uses dataset-specific parameters.
  • The approach gives a principled workflow: specify the selection diagram, identify an invariant conditional distribution, then choose a hierarchy whose levels mirror the invariant and variant components.
  • The licensing conditions from Theorem 1 can be used proactively to decide whether adding a dataset to the model's source pool would help a given subgroup.

Reading between the lines

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

  • Going beyond the paper, if the invariant demographic component P(Y|D) is as portable as the paper assumes, the same architecture should transfer to other diseases with stable demographic risk strata, such as tuberculosis or vaccine-preventable respiratory infections, whenever symptom-reporting mechanisms differ across data sources.
  • A direct stress test would simulate a target environment whose age- and gender-specific infection rates are deliberately shifted, violating S ⊥ Y | D; the model should degrade toward target-only performance, revealing how much of its gain comes from the invariant component versus from the hierarchy's shrinkage alone.
  • The subgroup-weight mechanism could be adapted as a diagnostic for sampling needs: subgroups whose learned weight falls mostly on θ_d are borrowing knowledge from other environments, which tells a surveillance program which populations are least represented locally.
  • The paper's framing suggests the same hierarchy could be coupled with fairness constraints at the subgroup level, since it already produces subgroup-specific classifiers rather than one aggregate model.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes a population-aware hierarchical Bayesian domain adaptation method, Hier+pop, for predicting influenza infection when the target dataset is largely unlabeled and comes from a new environment and population. The method is motivated by a causal selection diagram that distinguishes environment-specific symptom-reporting mechanisms from population-shared demographic characteristics. The authors claim that the model improves AUC over baselines on four real-world influenza datasets, especially when only 20% of target labels are available, and they provide a theorem describing when invariant versus dataset-specific parameters should be used. The manuscript includes implementation code and a supplementary proof sketch.

Significance. If the claims hold, the paper addresses a practically important problem: transferring prediction models across health datasets with different population compositions and study designs. The explicit incorporation of demographic attributes into the hierarchy, the use of multiple parent nodes, and the evaluation on four distinct real-world datasets are strengths, as is the release of code. The conceptual framing in terms of selection diagrams is also valuable. However, the central causal justification for the invariant component is not convincingly established, and there are inconsistencies in the formal statements that currently limit the paper's contribution.

major comments (4)
  1. [Section 4.1, Assumption 1; Section 4.1, Figure 1b] Assumption 1 states that no system variable directly causes any selection variable, but the text immediately before introduces an edge from D to \tilde{S} and Figure 1b depicts D→\tilde{S}. Since D is a system variable and \tilde{S} is a selection variable, this is an internal contradiction. The assumption must either be restricted to S* or the selection graph modified.
  2. [Section 4.3, Figure 1b, Section 5] The d-separation claim S⊥Y|D, which is used to justify transferring P(Y|D) as invariant information, is contradicted by the study designs described in Section 5. In Goviral and Fluwatch, participation is symptom-triggered and specimen submission is self-selected; in Hongkong, participants are household contacts of confirmed cases, making selection depend on exposure and infection; Hutterite data come from a vaccination trial. Under these mechanisms there are unblocked paths between S and Y given D, so P(Y|D) need not be equal across environments. The predictive gains may reflect useful hierarchical pooling, but the causal claim that the model harnesses population-invariant information is not supported by the evidence presented.
  3. [Section 4.7, Theorem 1 and Appendix A] Theorem 1 as stated is internally inconsistent. The main text gives conditions δD<δpop and PDa,g(Y)-Ppopa,g(Y)≈1 for using θ_l, with the proof sketch arguing that in the first case 'the specific dataset has more information'. The appendix, however, analyzes δD>δpop and concludes ID<Ipop, which would support using θ_d rather than θ_l. The proof does not establish the theorem as stated, and the conditions in the theorem therefore cannot be used to interpret the parameter-selection results in Table 3.
  4. [Sections 6 and 7.1, Tables 2 and 3, Figure 4] AUC values are reported without error bars, confidence intervals, or significance tests, and the number of random splits or repetitions is not specified. The hyperparameters λ=1 and β=0.2 are said to be selected by tuning, but the tuning procedure and the sensitivity of results to these choices are not described. These omissions make it difficult to assess whether the reported improvements of Hier+pop over baselines are statistically robust or specific to the chosen configuration.
minor comments (5)
  1. [Table 2] The dataset name 'Hutterite' is misspelled as 'Huttterite' in the table header.
  2. [Figure 2 caption] The caption contains 'dender' instead of 'gender'.
  3. [Section 1, Contributions] The phrase 'showing significant improvement' is used in the contributions list, but no statistical significance testing is reported; consider replacing 'significant' with 'consistent' or add formal tests.
  4. [Section 4.4, Equation (1)] The objective function uses f_j as positive predictive values rather than a likelihood; the relationship between this objective and a proper Bayesian hierarchical model could be clarified in the text.
  5. [Section 4.6] The restriction that the learned weights γ are positive is mentioned without an explanation of how such constraints are enforced in the nonlinear least-squares optimization; a sentence describing the constrained optimization method would improve reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the invariance claim is an explicit modeling assumption and the reported gains are empirical fitting results.

full rationale

The paper's derivation chain is an explicit causal-modeling assumption followed by empirical-Bayes fitting. The transferable invariant P(Y|D) is not derived from data; it is assumed by constructing the selection diagram without S->Y or S->D edges and by stating Assumptions 1 and 2 (e.g., 'The absence of a selection variable pointing at D and Y indicates that the mechanism of assigning values to these variables is the same across environments'). That is a modeling assumption, contestable on correctness grounds, but not circular: the conclusion S⊥Y|D follows from the stated graph rather than being fitted to the outcome. The hierarchical parameters and the gamma combination weights are estimated from the labeled source and target training subsets, and evaluation is on the held-out target portion ('All methods are tested on the target dataset (excluding the subset of the target data used for training)'), so no fitted parameter is renamed as a prediction. The empirical-Bayes priors centered on cumulative data are standard practice and do not make the test result an input. The 'licensing conditions' theorem is presented as a post-hoc characterization of when the model prefers theta_l versus theta_d, not as a first-principles prediction of a constant equal to its own input. Self-citations to prior work on participatory influenza data are used as domain-knowledge support and as baselines, not as a load-bearing uniqueness theorem. No equation in the paper reduces by construction to its own inputs, so no specific circular step can be exhibited.

Assumptions & free parameters 5 free parameters · 6 assumptions · 0 invented entities

The central modeling claim rests on the assumed causal graph and on the invariance of P(Y|D). The free parameters are mostly regularization and combination weights; they do not reach the level of deriving constants from first principles.

free parameters (5)
  • lambda (smoothing regularizer) = 1
    Set by hand in Section 4.6 to allow Laplacian smoothing; controls the unary term in the objective.
  • beta (hierarchy influence weight) = 0.2
    Chosen by hyperparameter tuning in Section 4.6; controls how strongly child parameters are pulled toward their parents.
  • gamma weights (gamma0 to gamma3) = Estimated per subgroup
    Learned by non-linear least squares in Section 4.6 from labeled data; they determine how much weight is given to dataset-level, age-level, and gender-level parameters.
  • Empirical-Bayes prior centers = Computed from cumulative data
    Section 4.5 centers priors on data-based estimates rather than using vague priors; this injects data-dependent information into the posterior.
  • Hierarchy node parameters theta (dataset, collection, age, gender, root) = MAP estimates from Equation 1
    These are the fitted parameters of the transfer model; the central claim about improved prediction depends on their joint estimation.
assumptions (6)
  • domain assumption The data generating process is known and representable as the selection diagram in Figure 1b.
    Section 4.1 states the main assumption that the DGP can be represented graphically, citing Pearl and Bareinboim; the invariant components are read off this graph.
  • domain assumption S is independent of Y given D, so P(Y|D) is invariant and transportable across environments.
    Section 4.3 derives this by d-separation in Figure 1b; if the graph is wrong, the transferred demographic component is wrong.
  • domain assumption The mechanism assigning Y, lab-confirmed influenza infection, is the same in all environments.
    Assumption 2 and the text around Figure 1b state that no selection variable points to Y; different lab or case-ascertainment practices would break this.
  • standard math P(V) is Markov and faithful with respect to G.
    Assumption 2(1); this is the standard condition connecting d-separation to conditional independence.
  • domain assumption Age and gender groups share stable infection characteristics across environments.
    Sections 1 and 4.3 invoke health literature that susceptibility and symptom patterns vary by age and gender; the model treats these as invariant.
  • ad hoc to paper Empirical-Bayes priors centered on cumulative data are weakly informative and do not leak test information.
    Section 4.5 sets priors from data without describing a validation or leakage-protection protocol; this is a modeling choice specific to the paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Population-aware Hierarchical Bayesian Domain Adaptation via Multiple-component Invariant Learning." pith.science (2026). https://pith.science/paper/3WDRZPV2

@misc{pith2026190809222,
  author       = {Pith},
  title        = {Pith review of: Population-aware Hierarchical Bayesian Domain Adaptation via Multiple-component Invariant Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3WDRZPV2}},
  note         = {Machine review of arXiv:1908.09222}
}
read the original abstract

While machine learning is rapidly being developed and deployed in health settings such as influenza prediction, there are critical challenges in using data from one environment in another due to variability in features; even within disease labels there can be differences (e.g. "fever" may mean something different reported in a doctor's office versus in an online app). Moreover, models are often built on passive, observational data which contain different distributions of population subgroups (e.g. men or women). Thus, there are two forms of instability between environments in this observational transport problem. We first harness knowledge from health to conceptualize the underlying causal structure of this problem in a health outcome prediction task. Based on sources of stability in the model, we posit that for human-sourced data and health prediction tasks we can combine environment and population information in a novel population-aware hierarchical Bayesian domain adaptation framework that harnesses multiple invariant components through population attributes when needed. We study the conditions under which invariant learning fails, leading to reliance on the environment-specific attributes. Experimental results for an influenza prediction task on four datasets gathered from different contexts show the model can improve prediction in the case of largely unlabelled target data from a new environment and different constituent population, by harnessing both environment and population invariant information. This work represents a novel, principled way to address a critical challenge by blending domain (health) knowledge and algorithmic innovation. The proposed approach will have a significant impact in many social settings wherein who and where the data comes from matters.

Figures

Figures reproduced from arXiv: 1908.09222 by the authors.

Figure 1
Figure 1. (a) Causal diagram, (b) Selection diagram represent [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Population-aware hierarchical model; θ parame￾ters at different nodes, D different data sets, α the priors. (A): Root level that represents invariant information across all data, (B): population parameters and information invariant to population-attributes (age) and (дender), (C): data set and environment-specific parameters and information (csfor cit￾izen science and hw for healthworker facilitated datasets). P (X,… view at source ↗
Figure 3
Figure 3. Demographic distributions in datasets. Darker [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Performance of Hier+pop method in comparison with baseline methods across increasing proportion of labelled [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 32 canonical work pages

  1. [1]

    Bansal, B

    S. Bansal, B. Pourbohloul, N. Hupert, B. Grenfell, and L. A. Meyers. The shifting demographic landscape of pandemic influenza. PLoS One, 5(2):e9360, 2010

    work page 2010

  2. [2]

    B. P. Carlin and T. A. Louis. Bayes and empirical Bayes methods for data analysis . Chapman and Hall/CRC, 2010

    work page 2010

  3. [3]

    Flu symptoms and diagnosis

    CDC. Flu symptoms and diagnosis. https://www.cdc.gov/flu/symptoms/index. html, 2019

    work page 2019

  4. [4]

    I. Chen, F. D. Johansson, and D. Sontag. Why is my classifier discriminatory? In Advances in Neural Information Processing Systems , pages 3539–3550, 2018

    work page 2018

  5. [5]

    Chunara, E

    R. Chunara, E. Goldstein, O. Patterson-Lomba, and J. S. Brownstein. Estimating influenza attack rates in the united states using a participatory cohort. Scientific reports, 5:9540, 2015

    work page 2015

  6. [6]

    B. J. Cowling, K. H. Chan, V. J. Fang, L. L. Lau, H. C. So, R. O. Fung, E. S. Ma, A. S. Kwong, C.-W. Chan, W. W. Tsui, et al. Comparative epidemiology of pandemic and seasonal influenza a in households. NEJM, 362(23):2175–2184, 2010

    work page 2010

  7. [7]

    W. F. Darnieder. Bayesian methods for data-dependent priors . PhD thesis, The Ohio State University, 2011

    work page 2011

  8. [8]

    Daume III

    H. Daume III. Frustratingly easy domain adaptation. arXiv preprint arXiv:0907.1815, 2009

    arXiv 2009

Show all 37 references
  1. [9]

    Elidan, B

    G. Elidan, B. Packer, G. Heitz, and D. Koller. Convex point estimation using undirected bayesian transfer hierarchies. arXiv preprint arXiv:1206.3252, 2012

    2012 arXiv

  2. [10]

    C. R. Evans, D. R. Williams, J.-P. Onnela, and S. Subramanian. A multilevel approach to modeling health inequalities at the intersection of multiple social identities. Social Science & Medicine , 203:64–73, 2018

    2018

  3. [11]

    J. R. Finkel and C. D. Manning. Hierarchical bayesian domain adaptation. In NAACL HLT, pages 602–610, 2009

    2009

  4. [12]

    Fletcher and M

    R. Fletcher and M. J. Powell. A rapidly convergent descent method for minimiza- tion. The computer journal, 6(2):163–168, 1963

    1963

  5. [13]

    E. B. Fragaszy, C. Warren-Gash, L. Wang, A. Copas, O. Dukes, W. J. Edmunds, N. Goonetilleke, G. Harvey, A. M. Johnson, J. Kovar, et al. Cohort profile: The flu watch study. International journal of epidemiology , 46(2):e18–e18, 2016

    2016

  6. [14]

    J. Goff, A. Rowe, J. S. Brownstein, and R. Chunara. Surveillance of acute res- piratory infections using community-submitted symptoms and specimens for molecular diagnostic testing. PLoS currents, 7, 2015

    2015

  7. [15]

    J. Guo, D. J. Shah, and R. Barzilay. Multi-source domain adaptation with mixture of experts. arXiv preprint arXiv:1809.02256, 2018

    2018 arXiv

  8. [16]

    M. Loeb, M. L. Russell, L. Moss, K. Fonseca, J. Fox, D. J. Earn, F. Aoki, G. Horsman, P. Van Caeseele, K. Chokani, et al. Effect of influenza vaccination of children on infection rates in hutterite communities: a randomized trial. JAMA, 303(10): 943–950, 2010

    2010

  9. [17]

    Magliacane, T

    S. Magliacane, T. van Ommen, T. Claassen, S. Bongers, P. Versteeg, and J. M. Mooij. Domain adaptation by using causal inference to predict invariant conditional distributions. In Advances in Neural Information Processing Systems , pages 10869– 10879, 2018

    2018

  10. [18]

    A. S. Monto, S. Gravenstein, M. Elliott, M. Colopy, and J. Schweinle. Clinical signs and symptoms predicting influenza infection. Archives of internal medicine, 160(21):3243–3247, 2000

    2000

  11. [19]

    J. M. Mooij, S. Magliacane, and T. Claassen. Joint causal inference from multiple contexts. arXiv preprint arXiv:1611.10351, 2016

    2016 arXiv

  12. [20]

    Mullainathan and Z

    S. Mullainathan and Z. Obermeyer. Who is tested for heart attack and who should be: Predicting patient risk and physician error. Technical report, National Bureau of Economic Research, 2019. Population-aware Hierarchical Bayesian Domain Adaptation via Multi-component Invariant...

    2019

  13. [21]

    J. Pearl. Causality: models, reasoning, and inference. IIE Transactions, 34(6): 583–589, 2002

    2002

  14. [22]

    Pearl and E

    J. Pearl and E. Bareinboim. Transportability of causal and statistical relations: A formal approach. In Twenty-Fifth AAAI Conference on Artificial Intelligence , 2011

    2011

  15. [23]

    A. L. Pineda, Y. Ye, S. Visweswaran, G. F. Cooper, M. M. Wagner, and F. R. Tsui. Comparison of machine learning classifiers for influenza detection from emer- gency department free-text reports. JBI, 58:60–69, 2015

    2015

  16. [24]

    Pivovarov, D

    R. Pivovarov, D. J. Albers, G. Hripcsak, J. L. Sepulveda, and N. Elhadad. Temporal trends of hemoglobin a1c testing. JAMIA, 21(6):1038–1044, 2014

    2014

  17. [25]

    Ray and R

    B. Ray and R. Chunara. Predicting acute respiratory infections from participatory data. Online journal of public health informatics , 9(1), 2017

    2017

  18. [26]

    Rehman, M

    N. Rehman, M. M. Aliapoulios, D. Umarwani, and R. Chunara. Domain adaptation for infection prediction from symptoms based on data from different study designs and contexts. arXiv preprint arXiv:1806.08835, 2018

    2018 arXiv

  19. [27]

    N. G. Reich, L. C. Brooks, S. J. Fox, S. Kandula, C. J. McGowan, E. Moore, D. Osthus, E. L. Ray, A. Tushar, T. K. Yamana, et al. A collaborative multiyear, multimodel assessment of seasonal influenza forecasting in the united states. Proceedings of the National Academy of Scie...

    2019

  20. [28]

    Saria, D

    S. Saria, D. Koller, and A. Penn. Learning individual and population level traits from clinical temporal data. In Proceedings of Neural Information Processing Systems, 2010

    2010

  21. [29]

    Singh, R

    H. Singh, R. Singh, V. Mhasawade, and R. Chunara. Fair predictors under distri- bution shift. arXiv preprint arXiv:1911.00677, 2019

    1911 arXiv

  22. [30]

    Subbaswamy, P

    A. Subbaswamy, P. Schulam, and S. Saria. Learning predictive models that transport. arXiv:1812.04597, 2018

    2018 arXiv

  23. [31]

    Suresh and J

    H. Suresh and J. V. Guttag. A framework for understanding unintended consequences of machine learning. CoRR, abs/1901.10002, 2019. URL http: //arxiv.org/abs/1901.10002

    1901 arXiv

  24. [32]

    Susser and E

    M. Susser and E. Susser. Choosing a future for epidemiology: Ii. from black box to chinese boxes and eco-epidemiology. American journal of public health , 86(5): 674–677, 1996

    1996

  25. [33]

    Ustun, Y

    B. Ustun, Y. Liu, and D. Parkes. Fairness without harm: Decoupled classifiers with preference guarantees. In International Conference on Machine Learning , pages 6373–6382, 2019

    2019

  26. [34]

    van Erp, J

    S. van Erp, J. Mulder, and D. L. Oberski. Prior sensitivity analysis in default bayesian structural equation modeling. American Psychological Association, 2017

    2017

  27. [35]

    Wiens, J

    J. Wiens, J. Guttag, and E. Horvitz. A study in transfer learning: leveraging data from multiple hospitals to enhance hospital-specific predictions. JAMIA, 21(4): 699–706, 2014

    2014

  28. [36]

    Influenza fact sheet: Overview

    World Health Organization. Influenza fact sheet: Overview. 2018

    2018

  29. [37]

    H. H. Zhou, V. Singh, S. C. Johnson, G. Wahba, A. D. N. Initiative, et al. Statistical tests and identifiability conditions for pooling and analyzing multisite datasets. PNAS, 115(7):1481–1486, 2018. ACM CHIL ’20, April 2–4, 2020, Toronto, ON, Canada Vishwali Mhasawade, Nabeel...

    2018

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.