pith. sign in

arxiv: 2605.20468 · v1 · pith:PZFODFCXnew · submitted 2026-05-19 · 💻 cs.LG · stat.ME· stat.ML

CASCADE Conformal Prediction: Uncertainty-Adaptive Prediction Intervals for Two-Stage Clinical Decision Support

Ricardo Diaz-Rincon , Muxuan Liang , Adolfo Ramirez-Zamora , Benjamin Shickel This is my paper

Pith reviewed 2026-05-21 07:12 UTC · model grok-4.3

classification 💻 cs.LG stat.MEstat.ML
keywords conformal predictionprediction intervalsepistemic uncertaintyParkinson's diseaseclinical decision supportmedication dosingVenn-Abersadaptive uncertainty
0
0 comments X

The pith

Mapping classifier uncertainty to non-conformity scores creates adaptive prediction intervals for Parkinson's medication dosing.

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

The paper introduces a two-stage approach for forecasting medication dose changes in Parkinson's disease. A classifier first identifies whether a change is needed and supplies its epistemic uncertainty, which is then mapped to adjust the size of prediction intervals produced by a regression model for the amount of change. This produces tighter intervals when the classifier is confident and wider intervals when uncertainty is high, aiming for better efficiency and maintained coverage compared to fixed conformal methods. A sympathetic reader would care because clinical AI often fails to signal varying reliability, and tailored intervals could help doctors act more precisely on clear cases while exercising caution on ambiguous ones.

Core claim

The central claim is that epistemic uncertainty from a primary classification task identifying medication change needs can be mapped directly to non-conformity scores for a secondary regression task predicting the size of that change. Implemented as CASCADE, this mapping enables continuous risk adaptation so that prediction intervals narrow for high-confidence patients and expand automatically for uncertain ones, yielding intervals 38.9 percent narrower than standard conformal baselines while preserving coverage.

What carries the argument

Direct mapping of Venn-Abers multi-probabilistic uncertainty from the screening classifier to non-conformity scores, which scales the regression prediction intervals without requiring auxiliary residual models.

If this is right

  • Intervals become narrower for patients where the classifier confidently detects no medication change needed
  • Intervals expand for high-uncertainty cases to preserve valid coverage
  • The method links discrete classification decisions directly to continuous dose regression forecasts
  • It removes the requirement for separate residual regression to quantify uncertainty

Where Pith is reading between the lines

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

  • The cascade mapping could be tested on sequential tasks in other chronic conditions such as diabetes management or heart failure titration
  • If coverage holds, clinicians might receive more actionable reliability signals that align interval width with decision confidence
  • The approach might combine with additional uncertainty sources to further refine adaptation in multi-stage pipelines

Load-bearing premise

Epistemic uncertainty estimates from the primary screening classifier can be directly mapped to non-conformity scores for the secondary regression task without breaking the marginal coverage guarantees of conformal prediction.

What would settle it

A held-out validation set where the observed coverage rate for the adaptive intervals drops below the nominal target specifically among cases the classifier flags as uncertain would falsify the central claim.

Figures

Figures reproduced from arXiv: 2605.20468 by Adolfo Ramirez-Zamora, Benjamin Shickel, Muxuan Liang, Ricardo Diaz-Rincon.

Figure 1
Figure 1. Figure 1: Overview of the CASCADE Framework. (a) Two-Stage Architecture: A Stage 1 classifier identifies patients requiring medication adjustment, followed by a Stage 2 regressor that predicts the dosage. (b) Uncertainty Quantification: CASCADE extracts the epistemic uncertainty score uVA(x) directly from the Stage 1 predicted probability pˆ(x). This score flows into the scaling function σ(x) (highlighted in red), w… view at source ↗
Figure 2
Figure 2. Figure 2: The CASCADE Mechanism. (a) Standard System: The regression model is blind to upstream confidence, producing non-informative, fixed-length prediction intervals for both the low uncertainty Patient A (p = 0.99) and the high uncertainty (ambiguous) Patient B (p = 0.55). While actionable for A, this rigidity is dangerous for B: the model confidently predicts a dosage increase despite high ambiguity. This erron… view at source ↗
Figure 3
Figure 3. Figure 3: β Ablation Study. The trade-off between safety and adaptivity. The dark red line denotes nominal coverage (target 80%), while the orange line denotes the Cascade Ratio (adaptivity). The optimal parameter β = 0.7 (gold star) maximizes adaptivity (CR 4.23) while strictly maintaining adequate coverage (80.1%). Lower values are inefficient while higher values violate coverage. While extreme scaling (β = 1.2) r… view at source ↗
read the original abstract

Effective medication management in Parkinson's Disease (PD) is challenging due to heterogeneous disease progression, variable patient response, and medication side effects. While AI models can forecast levodopa equivalent daily dose (LEDD) as a measure of medication needs, standard uncertainty quantification often fails to communicate the reliability of these predictions, treating high and low confidence clinical decisions identically. We introduce CASCADE (Calibrated Adaptive Scaling via Conformal And Distributional Estimation), a novel conformal prediction framework that propagates epistemic uncertainty from a screening classifier to adapt downstream predictions. Unlike standard conformal methods that rely on auxiliary residual regression, we leverage epistemic uncertainty from a primary classification task (identifying whether a medication change is needed) to dynamically scale the prediction intervals of a secondary regression task (predicting how much change). By mapping Venn-Abers multi-probabilistic uncertainty directly to non-conformity scores, our framework achieves continuous risk adaptation. We demonstrate that this ``cascade effect'' produces highly efficient intervals for confident patients (38.9% narrower than standard conformal baselines) while automatically expanding intervals to ensure robust coverage for uncertain cases, bridging the gap between discrete clinical decision-making and continuous dose forecasting in PD.

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 / 2 minor

Summary. The paper introduces CASCADE, a conformal prediction framework for two-stage clinical decision support in Parkinson's Disease medication management. It propagates epistemic uncertainty from a primary Venn-Abers classifier (identifying whether a medication change is needed) to dynamically adapt the prediction intervals of a secondary regression task (forecasting the levodopa equivalent daily dose change) via direct mapping of multi-probabilistic uncertainty to non-conformity scores. This is claimed to achieve continuous risk adaptation, yielding 38.9% narrower intervals for confident patients while automatically expanding intervals to maintain robust coverage for uncertain cases.

Significance. If the proposed mapping preserves marginal coverage, the work offers a novel way to cascade uncertainty across classification and regression stages without auxiliary residual models, which could improve efficiency in clinical AI applications where decision reliability varies. The explicit construction rather than parameter fitting is a methodological strength that aligns with conformal prediction principles.

major comments (2)
  1. [Abstract / Proposed Framework] The central claim rests on the validity of directly mapping Venn-Abers epistemic uncertainty to non-conformity scores for the regression task while retaining marginal coverage P(y in C(x)) >= 1-alpha. Standard conformal guarantees derive from exchangeability and calibration-set quantiles; the abstract describes the mapping as 'direct' but provides no derivation showing that the transformed scores preserve uniform p-value distribution under the null or that the adaptive intervals satisfy the coverage guarantee marginally. This is load-bearing, as efficiency gains could come at the cost of undercovering in high-uncertainty regimes.
  2. [Abstract / Experimental Results] The abstract asserts a 38.9% narrower interval width relative to standard conformal baselines and robust coverage for uncertain cases, yet the provided text contains no experimental details, coverage plots, ablation results, or quantitative validation across uncertainty strata. Without these, the performance claims cannot be verified and the weakest assumption (valid direct mapping without breaking guarantees) remains untested.
minor comments (2)
  1. [Notation / Method] Formalize the exact mapping from Venn-Abers probabilities to the non-conformity score with an equation; clarify whether the transformation is monotonic and how it interacts with the quantile estimation step.
  2. [Introduction / Experiments] Include a brief description of the PD dataset, feature sets for the two stages, and the specific alpha levels used to make the clinical setup reproducible.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive and insightful comments on our manuscript introducing the CASCADE framework. We have carefully addressed each major comment below with clarifications on the theoretical foundations and experimental support. Revisions have been made to strengthen the derivation of coverage guarantees and to better highlight the empirical validations.

read point-by-point responses

  1. Referee: [Abstract / Proposed Framework] The central claim rests on the validity of directly mapping Venn-Abers epistemic uncertainty to non-conformity scores for the regression task while retaining marginal coverage P(y in C(x)) >= 1-alpha. Standard conformal guarantees derive from exchangeability and calibration-set quantiles; the abstract describes the mapping as 'direct' but provides no derivation showing that the transformed scores preserve uniform p-value distribution under the null or that the adaptive intervals satisfy the coverage guarantee marginally. This is load-bearing, as efficiency gains could come at the cost of undercovering in high-uncertainty regimes.

    Authors: We appreciate the referee highlighting the importance of rigorously establishing the coverage property. Section 3 of the manuscript defines the direct mapping from Venn-Abers probabilities (p0, p1) to non-conformity scores via a monotonic scaling of the residual by the epistemic uncertainty measure (1 - |p0 - p1|). In the revised version, we have added a formal derivation (new Appendix A) showing that because the mapping is a continuous, strictly increasing transformation of the original calibration residuals, the rank-order statistics and thus the (1-α)-quantile thresholds are preserved. This ensures that the p-values remain uniformly distributed under the null of exchangeability, so the marginal coverage guarantee P(y ∈ C(x)) ≥ 1-α continues to hold by the standard conformal argument. The adaptive expansion for high-uncertainty cases is achieved by inflating the effective score, which widens rather than narrows the interval and therefore cannot produce undercovering. revision: yes

  2. Referee: [Abstract / Experimental Results] The abstract asserts a 38.9% narrower interval width relative to standard conformal baselines and robust coverage for uncertain cases, yet the provided text contains no experimental details, coverage plots, ablation results, or quantitative validation across uncertainty strata. Without these, the performance claims cannot be verified and the weakest assumption (valid direct mapping without breaking guarantees) remains untested.

    Authors: We apologize that the abstract did not explicitly reference the supporting experiments. The full manuscript contains these validations in Section 4: empirical coverage is reported stratified by uncertainty quantiles (Table 2), showing rates consistently above the nominal 1-α level; interval-width reductions of 38.9% are quantified for the lowest-uncertainty stratum with corresponding plots (Figure 3); and ablation experiments that disable the cascade mapping confirm both the efficiency gain and the necessity of uncertainty propagation for maintaining coverage. To address the concern, we have expanded the abstract with a concise summary of these results and added explicit cross-references to the figures and tables in the results section. revision: yes

Circularity Check

0 steps flagged

No significant circularity; central construction is an explicit proposed mapping

full rationale

The paper's core contribution is the explicit construction of CASCADE, which defines a direct mapping from Venn-Abers epistemic uncertainty (produced by the primary classifier) to scaled nonconformity scores for the secondary regression task. This mapping is introduced as a novel design choice rather than derived from or reduced to any fitted parameter, self-referential equation, or prior self-citation. The claimed efficiency gains (narrower intervals for confident cases) and coverage properties are presented as empirical outcomes of this construction, not as tautological consequences of the inputs. No load-bearing step reduces by definition to its own outputs, and the framework remains self-contained against external conformal prediction benchmarks without invoking unverified uniqueness theorems or ansatzes from the authors' prior work.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The method rests on the standard validity theorem of conformal prediction and on the assumption that classifier uncertainty can be treated as a valid non-conformity measure; no new free parameters or invented entities are introduced in the abstract.

axioms (1)
  • standard math Conformal prediction guarantees marginal coverage when non-conformity scores are exchangeable with the test point.
    Invoked implicitly when the paper claims robust coverage for uncertain cases.

pith-pipeline@v0.9.0 · 5752 in / 1249 out tokens · 53282 ms · 2026-05-21T07:12:14.463718+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

300 extracted references · 300 canonical work pages

  1. [1]

    Conformal Regression for Quantitative Structure-Activity Relationship Modeling - Quantifying Prediction Uncertainty , volume =

    Fredrik Svensson and Natalia Aniceto and Ulf Norinder and Isidro Cortes-Ciriano and Ola Spjuth and Lars Carlsson and Andreas Bender , doi =. Conformal Regression for Quantitative Structure-Activity Relationship Modeling - Quantifying Prediction Uncertainty , volume =. Journal of Chemical Information and Modeling , month =

    work page

  2. [2]

    Neurology , volume=

    Risk factors for progression in Parkinson's disease , author=. Neurology , volume=. 1988 , publisher=

    work page 1988

  3. [3]

    Parkinsonism & related disorders , volume=

    Classifying risk factors for dyskinesia in Parkinson’s disease , author=. Parkinsonism & related disorders , volume=. 2010 , publisher=

    work page 2010

  4. [4]

    Journal of neuroscience methods , volume=

    Non-invasive brain stimulation for Parkinson’s disease: Clinical evidence, latest concepts and future goals: A systematic review , author=. Journal of neuroscience methods , volume=. 2021 , publisher=

    work page 2021

  5. [5]

    Espay and Francesca Morgante and Aristide Merola and Alfonso Fasano and Luca Marsili and Susan H

    Alberto J. Espay and Francesca Morgante and Aristide Merola and Alfonso Fasano and Luca Marsili and Susan H. Fox and Erwan Bezard and Barbara Picconi and Paolo Calabresi and Anthony E. Lang , doi =. Levodopa-induced dyskinesia in Parkinson disease: Current and evolving concepts , volume =. Annals of Neurology , month =

    work page

  6. [6]

    Jaws-x: addressing efficiency bottlenecks of conformal prediction under standard and feedback covariate shift , year =

    Drew Prinster and Suchi Saria and Anqi Liu , booktitle =. Jaws-x: addressing efficiency bottlenecks of conformal prediction under standard and feedback covariate shift , year =

    work page

  7. [7]

    Conformal validity guarantees exist for any data distribution (and how to find them) , year =

    Drew Prinster and Samuel Stanton and Anqi Liu and Suchi Saria , journal =. Conformal validity guarantees exist for any data distribution (and how to find them) , year =

    work page

  8. [8]

    Theoretical foundations of conformal prediction , year =

    Anastasios N Angelopoulos and Rina Foygel Barber and Stephen Bates , journal =. Theoretical foundations of conformal prediction , year =

    work page

  9. [9]

    Cross-conformal predictive distributions , year =

    Vladimir Vovk and Ilia Nouretdinov and Valery Manokhin and Alexander Gammerman , booktitle =. Cross-conformal predictive distributions , year =

    work page

  10. [10]

    Cross-conformal predictors , volume =

    Vladimir Vovk , doi =. Cross-conformal predictors , volume =. Annals of Mathematics and Artificial Intelligence , keywords =

    work page

  11. [11]

    Confidence predictions for the diagnosis of acute abdominal pain , year =

    Harris Papadopoulos and Alex Gammerman and Volodya Vovk , booktitle =. Confidence predictions for the diagnosis of acute abdominal pain , year =

    work page

  12. [12]

    Reliable diagnosis of acute abdominal pain with conformal prediction , volume =

    Harris Papadopoulos and Alex Gammerman and Volodya Vovk , issue =. Reliable diagnosis of acute abdominal pain with conformal prediction , volume =. Engineering Intelligent Systems , pages =

    work page

  13. [13]

    Unbiased confidence measures for stroke risk estimation based on ultrasound carotid image analysis , volume =

    Harris Papadopoulos and Efthyvoulos Kyriacou and Andrew Nicolaides , doi =. Unbiased confidence measures for stroke risk estimation based on ultrasound carotid image analysis , volume =. Neural Computing and Applications , keywords =

    work page

  14. [14]

    Timms and John Sinclair and Usha Menon and Ian Jacobs and Alex Gammerman , doi =

    Dmitry Devetyarov and Ilia Nouretdinov and Brian Burford and Stephane Camuzeaux and Aleksandra Gentry-Maharaj and Ali Tiss and Celia Smith and Zhiyuan Luo and Alexey Chervonenkis and Rachel Hallett and Volodya Vovk and Mike Waterfield and Rainer Cramer and John F. Timms and John Sinclair and Usha Menon and Ian Jacobs and Alex Gammerman , doi =. Conformal ...

    work page

  15. [15]

    Alnemer and Lama Rajab and Ibrahim Aljarah , doi =

    Loai M. Alnemer and Lama Rajab and Ibrahim Aljarah , doi =. Conformal Prediction Technique to Predict Breast Cancer Survivability , volume =. International Journal of Advanced Science and Technology , month =

    work page

  16. [16]

    Conformal prediction under feedback covariate shift for biomolecular design , volume =

    Clara Fannjiang and Stephen Bates and Anastasios N Angelopoulos and Jennifer Listgarten and Michael I Jordan , issue =. Conformal prediction under feedback covariate shift for biomolecular design , volume =. Proceedings of the National Academy of Sciences , pages =

    work page

  17. [17]

    Distributional conformal prediction , volume =

    Victor Chernozhukov and Kaspar Wüthrich and Yinchu Zhu , issue =. Distributional conformal prediction , volume =. Proceedings of the National Academy of Sciences , pages =

    work page

  18. [18]

    Conformal prediction under covariate shift , volume =

    Ryan J Tibshirani and Rina Foygel Barber and Emmanuel Candes and Aaditya Ramdas , journal =. Conformal prediction under covariate shift , volume =

    work page

  19. [19]

    Discriminative jackknife: Quantifying uncertainty in deep learning via higher-order influence functions , year =

    Ahmed Alaa and Mihaela Van Der Schaar , booktitle =. Discriminative jackknife: Quantifying uncertainty in deep learning via higher-order influence functions , year =

    work page

  20. [20]

    Predictive inference is free with the jackknife+-after-bootstrap , volume =

    Byol Kim and Chen Xu and Rina Barber , journal =. Predictive inference is free with the jackknife+-after-bootstrap , volume =

    work page

  21. [21]

    Predictive inference with the jackknife+ , volume =

    Rina Foygel Barber and Emmanuel J Candes and Aaditya Ramdas and Ryan J Tibshirani , issue =. Predictive inference with the jackknife+ , volume =. The Annals of Statistics , pages =

    work page

  22. [22]

    Inductive confidence machines for regression , year =

    Harris Papadopoulos and Kostas Proedrou and Volodya Vovk and Alex Gammerman , booktitle =. Inductive confidence machines for regression , year =

    work page

  23. [23]

    Widnell and Suzanne Ostrowizki and Paulo Fontoura and F

    Kenneth Marek and Danna Jennings and Shirley Lasch and Andrew Siderowf and Caroline Tanner and Tanya Simuni and Chris Coffey and Karl Kieburtz and Emily Flagg and Sohini Chowdhury and Werner Poewe and Brit Mollenhauer and Todd Sherer and Mark Frasier and Claire Meunier and Alice Rudolph and Cindy Casaceli and John Seibyl and Susan Mendick and Norbert Schu...

    work page

  24. [24]

    Nahab and Ritesh A

    Jeremy Watts and Anahita Khojandi and Rama Vasudevan and Fatta B. Nahab and Ritesh A. Ramdhani , doi =. Improving medication regimen recommendation for parkinson’s disease using sensor technology , volume =. Sensors , keywords =

    work page

  25. [25]

    Shamir and Trygve Dolber and Angela M

    Reuben R. Shamir and Trygve Dolber and Angela M. Noecker and Benjamin L. Walter and Cameron C. McIntyre , doi =. Machine learning approach to optimizing combined stimulation and medication therapies for Parkinson's disease , volume =. Brain Stimulation , keywords =

    work page

  26. [26]

    Salmanpour and Mahdi Hosseinzadeh and Mahya Bakhtiyari and Mehdi Maghsudi and Arman Rahmim , doi =

    Mohammad R. Salmanpour and Mahdi Hosseinzadeh and Mahya Bakhtiyari and Mehdi Maghsudi and Arman Rahmim , doi =. Prediction of drug amount in Parkinson's disease using hybrid machine learning systems and radiomics features , volume =. International Journal of Imaging Systems and Technology , keywords =

    work page

  27. [27]

    A decision support system based on recurrent neural networks to predict medication dosage for patients with Parkinson's disease , volume =

    Atiye Riasi and Mehdi Delrobaei and Mehri Salari , doi =. A decision support system based on recurrent neural networks to predict medication dosage for patients with Parkinson's disease , volume =. Scientific Reports , keywords =

    work page

  28. [28]

    Schiess and Xiaoqian Jiang , doi =

    Yejin Kim and Jessika Suescun and Mya C. Schiess and Xiaoqian Jiang , doi =. Computational medication regimen for Parkinson’s disease using reinforcement learning , volume =. Scientific Reports , month =

    work page

  29. [29]

    M. S. Hema and K. Meena and R. Maheshprabhu and M. Nageswara Guptha and G. Prema Arokia Mary , doi =. Parkinson Disease Prediction and Drug Personalization Using Machine Learning Techniques , year =. Industrial Internet of Things , month =

    work page

  30. [30]

    Classification based levodopamine response prediction in parkinson’s disorder , volume =

    Aman Jatain and Shalini Bhaskar Bajaj and Ritika Agarwal and Haziq Rahat Bullah , doi =. Classification based levodopamine response prediction in parkinson’s disorder , volume =. Applied Artificial Intelligence , pages =

    work page

  31. [31]

    Machine learning with optimization to create medicine intake schedules for Parkinson’s disease patients , volume =

    Tomasz Gutowski and Ryszard Antkiewicz and Stanisław Szlufik , doi =. Machine learning with optimization to create medicine intake schedules for Parkinson’s disease patients , volume =. PLoS ONE , month =

    work page

  32. [32]

    Continuous dopamine-receptor treatment of Parkinson's disease: scientific rationale and clinical implications , volume =

    C Warren Olanow and Jose A Obeso and Fabrizio Stocchi , issue =. Continuous dopamine-receptor treatment of Parkinson's disease: scientific rationale and clinical implications , volume =. The Lancet Neurology , pages =

    work page

  33. [33]

    Molecular mechanisms of L-DOPA-induced dyskinesia , volume =

    Peter Jenner , doi =. Molecular mechanisms of L-DOPA-induced dyskinesia , volume =. Nature Reviews Neuroscience , month =

    work page

  34. [34]

    Aromatic amino acids and modification of parkinsonism , volume =

    George C Cotzias and Melvin H Van Woert and Lewis M Schiffer , issue =. Aromatic amino acids and modification of parkinsonism , volume =. New England Journal of Medicine , pages =

    work page

  35. [35]

    Conformalized Quantile Regression , volume =

    Yaniv Romano and Evan Patterson and Emmanuel Candes , editor =. Conformalized Quantile Regression , volume =. Advances in Neural Information Processing Systems , publisher =

    work page

  36. [36]

    Jordan , journal =

    Anastasios Angelopoulos and Stephen Bates and Jitendra Malik and Michael I. Jordan , journal =. Uncertainty Sets for Image Classifiers using Conformal Prediction , url =

    work page

  37. [37]

    Algorithmic Learning in a Random World , year =

    Vladimir Vovk and Alexander Gammerman and Glenn Shafer , publisher =. Algorithmic Learning in a Random World , year =

    work page

  38. [38]

    The need for uncertainty quantification in machine-assisted medical decision making , volume =

    Edmon Begoli and Tanmoy Bhattacharya and Dimitri Kusnezov , doi =. The need for uncertainty quantification in machine-assisted medical decision making , volume =. Nature Machine Intelligence , month =

    work page

  39. [39]

    Latourelle and Michael T

    Jeanne C. Latourelle and Michael T. Beste and Tiffany C. Hadzi and Robert E. Miller and Jacob N. Oppenheim and Matthew P. Valko and Diane M. Wuest and Bruce W. Church and Iya G. Khalil and Boris Hayete and Charles S. Venuto , doi =. Large-scale identification of clinical and genetic predictors of motor progression in patients with newly diagnosed Parkinso...

    work page

  40. [40]

    Entropy-Based Machine Learning Model for Fast Diagnosis and Monitoring of Parkinson's Disease , volume =

    Maksim Belyaev and Murugappan Murugappan and Andrei Velichko and Dmitry Korzun , doi =. Entropy-Based Machine Learning Model for Fast Diagnosis and Monitoring of Parkinson's Disease , volume =. Sensors (Basel, Switzerland) , keywords =

    work page

  41. [41]

    Fox and Regina Katzenschlager and Shen Yang Lim and Brandon Barton and Rob M.A

    Susan H. Fox and Regina Katzenschlager and Shen Yang Lim and Brandon Barton and Rob M.A. de Bie and Klaus Seppi and Miguel Coelho and Cristina Sampaio , doi =. International Parkinson and movement disorder society evidence-based medicine review: Update on treatments for the motor symptoms of Parkinson's disease , volume =. Movement Disorders , keywords =

    work page

  42. [42]

    Kalia and Ahsan Habib and Priya Jagota and Rajeev Ojha and Rukmini Mridula Kandadai and Soaham Desai and Manjula Caldera and Darshana Sirisena and Divyani Garg and Tiago A

    Prashanth Lingappa Kukkle and Lorraine V. Kalia and Ahsan Habib and Priya Jagota and Rajeev Ojha and Rukmini Mridula Kandadai and Soaham Desai and Manjula Caldera and Darshana Sirisena and Divyani Garg and Tiago A. Mestre and Rosy Neupane and Suppata Maytharakcheep and Kanyawat Sanyawut and Rupam Borgohain , doi =. Levodopa Equivalent Daily Dosage: Geogra...

    work page

  43. [43]

    Tomlinson and Rebecca Stowe and Smitaa Patel and Caroline Rick and Richard Gray and Carl E

    Claire L. Tomlinson and Rebecca Stowe and Smitaa Patel and Caroline Rick and Richard Gray and Carl E. Clarke , doi =. Systematic review of levodopa dose equivalency reporting in Parkinson's disease , volume =. Movement Disorders , keywords =

    work page

  44. [44]

    Anang and Véronique Latreille and Jean François Gagnon and Ronald B

    Seyed Mohammad Fereshtehnejad and Silvia Rios Romenets and Julius B.M. Anang and Véronique Latreille and Jean François Gagnon and Ronald B. Postuma , doi =. New clinical subtypes of Parkinson disease and their longitudinal progression a prospective cohort comparison with other phenotypes , volume =. JAMA Neurology , month =

    work page

  45. [45]

    Variable expression of Parkinson's disease: A base-line analysis of the DATATOP cohort , url =

    J Jankovic and ; M Mcdermott and ; J Carter and ; S Gauthier and ; C Goetz and ; L Golbe and ; S Huber and ; W Koller and ; C Olanow and ; I Shoulson and ; M Stern and ; C Tanner and ; W Weiner and Parkinson Study Group , journal =. Variable expression of Parkinson's disease: A base-line analysis of the DATATOP cohort , url =

    work page

  46. [46]

    Bloem and Camille Carroll and Jesse Cedarbaum and Joshua Cosman and David T

    Amit Khanna and Jamie Adams and Chrystalina Antoniades and Bastiaan R. Bloem and Camille Carroll and Jesse Cedarbaum and Joshua Cosman and David T. Dexter and Marissa F. Dockendorf and Jeremy Edgerton and Laura Gaetano and Erkuden Goikoetxea and Derek Hill and Fay Horak and Elena S. Izmailova and Tairmae Kangarloo and Dina Katabi and Catherine Kopil and M...

    work page

  47. [47]

    Espay and Patrik Brundin and Anthony E

    Alberto J. Espay and Patrik Brundin and Anthony E. Lang , doi =. Precision medicine for disease modification in Parkinson disease , volume =. Nature Reviews Neurology , month =

    work page

  48. [48]

    Severson and Lana M

    Kristen A. Severson and Lana M. Chahine and Luba A. Smolensky and Murtaza Dhuliawala and Mark Frasier and Kenney Ng and Soumya Ghosh and Jianying Hu , doi =. Discovery of Parkinson's disease states and disease progression modelling: a longitudinal data study using machine learning , volume =. The Lancet Digital Health , month =

    work page

  49. [49]

    Personalized Input-Output Hidden Markov Models for Disease Progression Modeling , volume =

    Kristen A Severson and Lana M Chahine and Luba Smolensky and Kenney Ng and Jianying Hu and Soumya Ghosh , journal =. Personalized Input-Output Hidden Markov Models for Disease Progression Modeling , volume =

    work page

  50. [50]

    Beck , doi =

    Caroline Pearson and Alex Hartzman and Dianne Munevar and Megan Feeney and Rachel Dolhun and Veronica Todaro and Sheera Rosenfeld and Allison Willis and James C. Beck , doi =. Care access and utilization among medicare beneficiaries living with Parkinson’s disease , volume =. npj Parkinson's Disease , month =

    work page

  51. [51]

    Clarke , doi =

    Vincent Low and Yoav Ben-Shlomo and Elena Coward and Suzanne Fletcher and Richard Walker and Carl E. Clarke , doi =. Measuring the burden and mortality of hospitalisation in Parkinson's disease: A cross-sectional analysis of the English Hospital Episodes Statistics database 2009-2013 , volume =. Parkinsonism and Related Disorders , keywords =

    work page 2009

  52. [52]

    Loftus and Benjamin Shickel and Matthew M

    Tyler J. Loftus and Benjamin Shickel and Matthew M. Ruppert and Jeremy A. Balch and Tezcan Ozrazgat-Baslanti and Patrick J. Tighe and Philip A. Efron and William R. Hogan and Parisa Rashidi and Gilbert R. Upchurch and Azra Bihorac , doi =. Uncertainty-aware deep learning in healthcare: A scoping review , volume =. PLOS Digital Health , month =

    work page

  53. [53]

    Louise Rushworth and Stephen Tisch , doi =

    Michal Lubomski and R. Louise Rushworth and Stephen Tisch , doi =. Hospitalisation and comorbidities in Parkinson's disease: A large Australian retrospective study , volume =. Journal of Neurology, Neurosurgery and Psychiatry , month =

    work page

  54. [54]

    Haaxma and Bastiaan R

    Charlotte A. Haaxma and Bastiaan R. Bloem and George F. Borm and Wim J.G. Oyen and Klaus L. Leenders and Silvia Eshuis and Jan Booij and Dean E. Dluzen and Martin W.I.M. Horstink , doi =. Gender differences in Parkinson's disease , volume =. Journal of Neurology, Neurosurgery and Psychiatry , month =

    work page

  55. [55]

    Gerlach and Ania Winogrodzka and Wim E.J

    Oliver H.H. Gerlach and Ania Winogrodzka and Wim E.J. Weber , doi =. Clinical problems in the hospitalized Parkinson's disease patient: Systematic review , volume =. Movement Disorders , keywords =

    work page

  56. [56]

    Dumurgier and C

    J. Dumurgier and C. Tzourio , doi =. Epidemiology of neurological diseases in older adults , volume =. Revue Neurologique , keywords =

    work page

  57. [57]

    Rodriguez and Rebecca N

    Viviana A. Rodriguez and Rebecca N. Mahon and Elisabeth Weiss and Nitai D. Mukhopadhyay , doi =. Two-Part Mixed Effects Mixture Model for Zero-Inflated Longitudinal Compositional Data , year =. Journal of the Indian Society for Probability and Statistics , keywords =

    work page

  58. [58]

    V. T. Farewell and D. L. Long and B. D.M. Tom and S. Yiu and L. Su , doi =. Two-part and related regression models for longitudinal data , volume =. Annual Review of Statistics and Its Application , keywords =

    work page

  59. [59]

    A two-part mixed-effects pattern-mixture model to handle zero-inflation and incompleteness in a longitudinal setting , volume =

    Antonello Maruotti , doi =. A two-part mixed-effects pattern-mixture model to handle zero-inflation and incompleteness in a longitudinal setting , volume =. Biometrical Journal , keywords =

    work page

  60. [60]

    Two-part regression models for longitudinal zero-inflated count data , year =

    Marco Alfo and Antonello Mauotti , doi =. Two-part regression models for longitudinal zero-inflated count data , year =. Canadian Journal of Statistics , pages =

    work page

  61. [61]

    A comparison of zero-inflated and hurdle models for modeling zero-inflated count data , volume =

    Cindy Xin Feng , doi =. A comparison of zero-inflated and hurdle models for modeling zero-inflated count data , volume =. Journal of Statistical Distributions and Applications , keywords =

    work page

  62. [62]

    Zero-inflated models for adjusting varying exposures: a cautionary note on the pitfalls of using offset , volume =

    Cindy Feng , doi =. Zero-inflated models for adjusting varying exposures: a cautionary note on the pitfalls of using offset , volume =. Journal of Applied Statistics , keywords =

    work page

  63. [63]

    Understanding the Under-Coverage Bias in Uncertainty Estimation , url =

    Yu Bai and Song Mei and Huan Wang and Caiming Xiong , journal =. Understanding the Under-Coverage Bias in Uncertainty Estimation , url =

    work page

  64. [64]

    Improving Output Uncertainty Estimation and Generalization in Deep Learning via Neural Network Gaussian Processes , url =

    Tomoharu Iwata and Zoubin Ghahramani , journal =. Improving Output Uncertainty Estimation and Generalization in Deep Learning via Neural Network Gaussian Processes , url =

    work page

  65. [65]

    Keener and Elena Moro , doi =

    Indu Subramanian and Soania Mathur and Annelien Oosterbaan and Richelle Flanagan and Adrienne M. Keener and Elena Moro , doi =. Unmet Needs of Women Living with Parkinson's Disease: Gaps and Controversies , volume =. Movement Disorders , keywords =

    work page

  66. [66]

    Verschuur and Sven R

    Constant V.M. Verschuur and Sven R. Suwijn and Judith A. Boel and Bart Post and Bas R. Bloem and Johannes J. van Hilten and Teus van Laar and Gerrit Tissingh and Alexander G. Munts and Guenther Deuschl and Anthony E. Lang and Marcel G.W. Dijkgraaf and Rob J. de Haan and Rob M.A. de Bie , doi =. Randomized Delayed-Start Trial of Levodopa in Parkinson’s Dis...

    work page

  67. [67]

    García-Ruiz and Javier Del Val and Ignacio Mahillo Fernández and Antonio Herranz , doi =

    Pedro J. García-Ruiz and Javier Del Val and Ignacio Mahillo Fernández and Antonio Herranz , doi =. What factors influence motor complications in parkinson disease? A 10-year prospective study , volume =. Clinical Neuropharmacology , keywords =

    work page

  68. [68]

    Malin and You Chen , doi =

    Xinmeng Zhang and Chao Yan and Cheng Gao and Bradley A. Malin and You Chen , doi =. Predicting Missing Values in Medical Data Via XGBoost Regression , volume =. Journal of Healthcare Informatics Research , keywords =

    work page

  69. [69]

    Conformal Prediction with Conditional Guarantees , year =

    Isaac Gibbs and John J Cherian and Emmanuel J Candès , journal =. Conformal Prediction with Conditional Guarantees , year =

    work page

  70. [70]

    M. S. Wisz and R. J. Hijmans and J. Li and A. T. Peterson and C. H. Graham and A. Guisan and J. Elith and M. Dudík and S. Ferrier and F. Huettmann and J. R. Leathwick and A. Lehmann and L. Lohmann and B. A. Loiselle and G. Manion and C. Moritz and M. Nakamura and Y. Nakazawa and J. Mc C. Overton and S. J. Phillips and K. S. Richardson and R. Scachetti-Per...

    work page

  71. [71]

    Evaluating the state of the art in missing data imputation for clinical data , volume =

    Yuan Luo , doi =. Evaluating the state of the art in missing data imputation for clinical data , volume =. Briefings in Bioinformatics , keywords =

    work page

  72. [72]

    Miao and Eduardo Rodriguez Almaraz and Madhumita Sushil and Atul J

    Nikita Mehandru and Brenda Y. Miao and Eduardo Rodriguez Almaraz and Madhumita Sushil and Atul J. Butte and Ahmed Alaa , doi =. Evaluating large language models as agents in the clinic , volume =. npj Digital Medicine , month =

    work page

  73. [73]

    LightGBM: A Highly Efficient Gradient Boosting Decision Tree , url =

    Guolin Ke and Qi Meng and Thomas Finley and Taifeng Wang and Wei Chen and Weidong Ma and Qiwei Ye and Tie-Yan Liu , booktitle =. LightGBM: A Highly Efficient Gradient Boosting Decision Tree , url =

    work page

  74. [74]

    Tabular data: Deep learning is not all you need , volume =

    Ravid Shwartz-Ziv and Amitai Armon , doi =. Tabular data: Deep learning is not all you need , volume =. Information Fusion , keywords =

    work page

  75. [75]

    Conway and Hamid A

    Ezekwesiri Michael Nwanosike and Barbara R. Conway and Hamid A. Merchant and Syed Shahzad Hasan , doi =. Potential applications and performance of machine learning techniques and algorithms in clinical practice: A systematic review , volume =. International Journal of Medical Informatics , keywords =

    work page

  76. [76]

    Helou and Deborah DiazGranados and Michael S

    Marieka A. Helou and Deborah DiazGranados and Michael S. Ryan and John W. Cyrus , doi =. Uncertainty in Decision Making in Medicine: A Scoping Review and Thematic Analysis of Conceptual Models , volume =. Academic Medicine , pages =

    work page

  77. [77]

    Mukund Sundararajan and Amir Najmi , title =

    work page

  78. [78]

    An Overview of Overfitting and its Solutions , volume =

    Xue Ying , doi =. An Overview of Overfitting and its Solutions , volume =. Journal of Physics: Conference Series , month =

    work page

  79. [79]

    Hyperparameter optimization for machine learning models based on Bayesian optimization , volume =

    Jia Wu and Xiu Yun Chen and Hao Zhang and Li Dong Xiong and Hang Lei and Si Hao Deng , doi =. Hyperparameter optimization for machine learning models based on Bayesian optimization , volume =. Journal of Electronic Science and Technology , keywords =

    work page

  80. [80]

    Heterogeneity of Variance in Clinical Data , volume =

    Robert J Grissom , doi =. Heterogeneity of Variance in Clinical Data , volume =. Journal of Consulting and Clinical Psychology , pages =

    work page

Showing first 80 references.