pith. sign in

arxiv: 2504.18544 · v3 · pith:FLLRY2C7new · submitted 2025-04-10 · 💻 cs.LG · cs.AI· cs.CY

Critical Challenges and Guidelines in Evaluating Synthetic Tabular Data: A Systematic Review

Nazia Nafis , Inaki Esnaola , Alvaro Martinez-Perez , Maria-Cruz Villa-Uriol , Venet Osmani This is my paper

Pith reviewed 2026-05-22 21:10 UTC · model grok-4.3

classification 💻 cs.LG cs.AIcs.CY
keywords synthetic tabular datahealth dataevaluation methodssystematic reviewdata generationreproducibilitydomain expertsprivacy preservation
0
0 comments X

The pith

A review of 134 studies finds no consensus on evaluating synthetic tabular health data and offers taxonomies plus guidelines to fix inconsistent practices.

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

This systematic review looks at how synthetic tabular data, especially for health applications, is generated and checked for quality. The authors screened over 2000 papers from the last decade and closely examined 134 of them, uncovering problems such as differing ideas on what counts as good evaluation, metrics used in uneven ways, little input from doctors or other experts, skimpy descriptions of the original datasets, and results that others cannot easily repeat. To address these issues, the paper groups generation and evaluation techniques into organized taxonomies and supplies step-by-step guidelines meant to make assessments more reliable and comparable. Synthetic data is meant to let researchers work around privacy rules while still advancing medical studies, yet without solid ways to judge its trustworthiness, the data risks producing flawed models or unsafe clinical insights.

Core claim

The review identifies key challenges in evaluating synthetic tabular health data, including the lack of consensus on evaluation methods, inconsistent application of evaluation metrics, limited involvement of domain experts, inadequate reporting of dataset characteristics, and limited reproducibility of results. In response, the authors provide a structured consolidation of synthetic data generation and evaluation methods into taxonomies, alongside practical guidelines to support more robust and standardised evaluation practices.

What carries the argument

The taxonomies classifying synthetic data generation techniques and evaluation approaches, which organize existing methods to reveal inconsistencies and guide standardized reporting.

If this is right

  • Researchers adopting the guidelines will produce evaluations that can be compared more directly across different synthetic data generators.
  • Increased involvement of domain experts in evaluation will raise the clinical usefulness of the resulting synthetic datasets.
  • Standardized reporting of dataset characteristics will make it easier for other teams to reproduce and build on published results.
  • The taxonomies will help new studies select appropriate generation and evaluation methods rather than defaulting to the most common but inconsistent ones.

Where Pith is reading between the lines

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

  • The same challenges and taxonomies could apply to synthetic data in non-health domains such as finance or social science where privacy is also a concern.
  • The guidelines might serve as a starting point for developing automated software tools that check adherence to recommended evaluation steps.

Load-bearing premise

The search strategy and inclusion criteria that selected 134 studies from 2067 papers accurately represent the field without major bias in which papers were chosen or how challenges were interpreted.

What would settle it

A follow-up systematic review using similar methods but different databases that finds widespread agreement on a single set of evaluation metrics across most studies.

Figures

Figures reproduced from arXiv: 2504.18544 by Alvaro Martinez-Perez, Inaki Esnaola, Maria-Cruz Villa-Uriol, Nazia Nafis, Venet Osmani.

Figure 1
Figure 1. Figure 1: There is an increasing lack of appropriate evaluation metrics (due to increasing difficul [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: (L) Breakdown of Evaluation Approaches into Direct vs Indirect Evaluation, and (R) Breakdown of Evaluation Methods into Quantitative vs Qualitative Methods Manuscript submitted to ACM [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: (L) Breakdown of Quantitative Evaluation Methods into Statistical and ML-based methods, and (R) Breakdown of Qualitative Evaluation Methods into Visualisation and Domain Expert Review The popularity of Statistical evaluation techniques remains high, however, a trend can be seen in publications using both Statistical and ML-based techniques together ( [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Popularity trend of Statistical and ML-based Evaluation methods over the last decade, as obtained from publications included [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: (L) Most popular Evaluation Metrics, and (R) Breakdown of whether the metric is author-defined or not Qualitative evaluation methods rely on subjective judgement and human interpretation to assess the quality of synthetic data. In the majority of the cases (71.3%), they take the form of Visual Inspection of the graphical representation of distributions of synthetic data ( [PITH_FULL_IMAGE:figures/full_fig… view at source ↗
Figure 7
Figure 7. Figure 7: (L) Breakdown of Generation Models into Probabilistic vs Mechanistic Models, and (R) Most popular Generation Models We also observed a growing divergence between Probabilistic and Mechanistic models, with Probabilistic Models increasingly being more frequently used (74.3%) and Mechanistic Models tending to be more referenced in older publications only ( [PITH_FULL_IMAGE:figures/full_fig_p006_7.png] view at source ↗
Figure 6
Figure 6. Figure 6: Sankey depicting the most popular Evaluation Metrics and the papers utilising them, as obtained from the publications [PITH_FULL_IMAGE:figures/full_fig_p007_6.png] view at source ↗
Figure 8
Figure 8. Figure 8: Popularity trend of Probabilistic and Mechanistic Generation Models over the last decade, as obtained from publications [PITH_FULL_IMAGE:figures/full_fig_p008_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Most common diseases ( grouped by their ICD-9 codes), as seen through the publica [PITH_FULL_IMAGE:figures/full_fig_p008_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Popularity trend of synthetic tabular health data over the last decade, as obtained from publications included in this review [PITH_FULL_IMAGE:figures/full_fig_p011_10.png] view at source ↗
read the original abstract

Generating synthetic tabular health data is challenging, and evaluating their quality is equally, if not more, complex. This systematic review highlights the critical importance of rigorous evaluation of synthetic health data to ensure reliability, clinical relevance, and appropriate use. From an initial identification of 2067 relevant papers published in the last ten years, 134 studies were selected for detailed analysis. Our review identifies key challenges, including lack of consensus on evaluation methods, inconsistent application of evaluation metrics, limited involvement of domain experts, inadequate reporting of dataset characteristics, and limited reproducibility of results. In response, we provide a structured consolidation of synthetic data generation and evaluation methods into taxonomies, alongside practical guidelines to support more robust and standardised evaluation practices. These findings aim to support the responsible development and use of synthetic health data, aligned with emerging expectations around transparency, reproducibility, and governance, ultimately enabling the community to fully harness its transformative potential and accelerate innovation.

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. This systematic review examines evaluation practices for synthetic tabular health data. From 2067 papers identified over the last decade, the authors select and analyze 134 studies. They report five key challenges: lack of consensus on evaluation methods, inconsistent metric use, limited domain-expert involvement, inadequate reporting of dataset characteristics, and limited reproducibility. In response, the paper consolidates generation and evaluation methods into taxonomies and supplies practical guidelines intended to promote standardized, transparent evaluation.

Significance. If the selection of 134 studies is representative and the challenge extraction is reproducible, the taxonomies and guidelines would provide a useful consolidation for the field, directly addressing gaps in standardization and reproducibility. The systematic-review framing and explicit provision of structured taxonomies constitute a clear strength.

major comments (2)
  1. [Methods] Methods section (search and selection procedure): the manuscript states that 2067 papers were identified and 134 selected but supplies no explicit search strategy, database list, search strings, inclusion/exclusion criteria, or PRISMA-style flow diagram. Because the central claims rest on the representativeness of these 134 studies and the reproducibility of the identified challenges, this omission is load-bearing.
  2. [Results] Results section (challenge extraction): the paper lists five challenges but does not describe the coding scheme, inter-rater process, or how many papers contributed evidence to each challenge. Without this, it is impossible to assess whether the listed challenges are exhaustive or whether selection bias in the 134 studies could have produced them.
minor comments (2)
  1. [Abstract] Abstract: the sentence reporting the 2067-to-134 reduction should be accompanied by a parenthetical reference to the Methods section so readers immediately know where the protocol is described.
  2. [Taxonomies] Taxonomy figures: the generation and evaluation taxonomies would benefit from explicit cross-references in the text to the specific tables or figures that present them.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive and detailed feedback. We agree that greater methodological transparency is required to support the reproducibility and credibility of our systematic review. We address each major comment below and will revise the manuscript accordingly.

read point-by-point responses

  1. Referee: [Methods] Methods section (search and selection procedure): the manuscript states that 2067 papers were identified and 134 selected but supplies no explicit search strategy, database list, search strings, inclusion/exclusion criteria, or PRISMA-style flow diagram. Because the central claims rest on the representativeness of these 134 studies and the reproducibility of the identified challenges, this omission is load-bearing.

    Authors: We agree that the current Methods section provides only a high-level summary and lacks the necessary details. In the revised manuscript we will expand this section to include: the complete list of databases searched (PubMed, Scopus, Web of Science, IEEE Xplore, ACM Digital Library, and arXiv), the exact search strings and Boolean operators used, the full inclusion/exclusion criteria applied at title/abstract and full-text stages, and a PRISMA flow diagram documenting the screening process from 2067 records to the final 134 studies. These additions will directly substantiate the representativeness of the selected corpus and enable independent reproduction of the study selection. revision: yes

  2. Referee: [Results] Results section (challenge extraction): the paper lists five challenges but does not describe the coding scheme, inter-rater process, or how many papers contributed evidence to each challenge. Without this, it is impossible to assess whether the listed challenges are exhaustive or whether selection bias in the 134 studies could have produced them.

    Authors: We acknowledge this limitation in transparency. The challenges were derived via thematic analysis of the 134 papers, yet the manuscript does not report the coding protocol. We will revise the Results (and/or Methods) section to describe the coding scheme and thematic framework employed, clarify whether coding was conducted by one or multiple reviewers with discrepancy resolution, and report the number (or proportion) of papers providing evidence for each of the five challenges. This information will allow readers to evaluate exhaustiveness and potential selection bias. revision: yes

Circularity Check

0 steps flagged

No significant circularity; review synthesizes external literature without self-referential reductions.

full rationale

This systematic review derives its claims (challenges, taxonomies, guidelines) from analysis of 134 externally selected papers. No equations, fitted parameters, or predictions exist that reduce to the authors' own inputs by construction. No self-citation chains are load-bearing for the core findings. The search/inclusion process is a methodological step whose validity is separate from circularity as defined (no self-definitional, fitted-input, or uniqueness-imported patterns apply).

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the validity of standard systematic review methodology for literature selection and synthesis, with no free parameters, invented entities, or ad hoc axioms beyond domain assumptions about review practices.

axioms (1)
  • domain assumption Systematic reviews following established protocols provide a comprehensive and unbiased overview of the literature on synthetic data evaluation.
    Invoked implicitly in the selection of 134 studies from 2067 papers and the identification of challenges.

pith-pipeline@v0.9.0 · 5709 in / 1442 out tokens · 40775 ms · 2026-05-22T21:10:13.847959+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

121 extracted references · 121 canonical work pages

  1. [1]

    Ngo, Taridzo Chomutare, Hercules Dalianis, Elisa Salvi, Andrius Budrionis, and Fred Godtliebsen

    Maryam Tayefi, Phuong D. Ngo, Taridzo Chomutare, Hercules Dalianis, Elisa Salvi, Andrius Budrionis, and Fred Godtliebsen. Challenges and opportunities beyond structured data in analysis of electronic health records.Wiley Interdisciplinary Reviews: Computational Statistics, 13, 2021

    work page 2021

  2. [2]

    Data silos undermine efforts to characterize, predict, and mitigate dementia-related missing person incidents.Healthcare Management Forum, 35:084047042211061, 06 2022

    Antonio Cruz, Samantha Marshall, Christine Daum, Hector Perez, John Hirdes, and Lili Liu. Data silos undermine efforts to characterize, predict, and mitigate dementia-related missing person incidents.Healthcare Management Forum, 35:084047042211061, 06 2022

    work page 2022

  3. [3]

    From biobank and data silos into a data commons: convergence to support translational medicine, 08 2021

    Rebecca Asiimwe, Stephanie Lam, Samuel Leung, Shanzhao Wang, Rachel Wan, Anna Tinker, Jessica Mcalpine, Michelle Woo, David Huntsman, and Aline Talhouk. From biobank and data silos into a data commons: convergence to support translational medicine, 08 2021

    work page 2021

  4. [4]

    Data democratization: toward a deeper understanding

    Hippolyte Lefebvre, Christine Legner, and Martin Fadler. Data democratization: toward a deeper understanding. 09 2021

    work page 2021

  5. [5]

    Data quality in health research: an integrative literature review (preprint).Journal of Medical Internet Research, 25, 08 2022

    Filipe Bernardi, Domingos Alves, Nathalia Crepaldi, Diego Yamada, Vinícius Lima, and Rui Rijo. Data quality in health research: an integrative literature review (preprint).Journal of Medical Internet Research, 25, 08 2022

    work page 2022

  6. [6]

    Protecting privacy while optimizing the use of (health)data: The importance of measures and safeguards.The American Journal of Bioethics, 22:79–81, 07 2022

    Julie-Anne Smit, Menno Mostert, and Johannes Delden. Protecting privacy while optimizing the use of (health)data: The importance of measures and safeguards.The American Journal of Bioethics, 22:79–81, 07 2022

    work page 2022

  7. [7]

    Health city: Transforming health and driving economic development.Healthcare Management Forum, 34:084047042094226, 08 2020

    Reg Joseph, Antonio Bruni, and Chris Carvalho. Health city: Transforming health and driving economic development.Healthcare Management Forum, 34:084047042094226, 08 2020

    work page 2020

  8. [8]

    Synthetic data in medical research.BMJ Medicine, 1, 09 2022

    Theodora Kokosi and Katie Harron. Synthetic data in medical research.BMJ Medicine, 1, 09 2022

    work page 2022

  9. [9]

    Harnessing the power of synthetic data in healthcare: innovation, application, and privacy.npj Digital Medicine, 6, 10 2023

    Mauro Giuffrè and Dennis Shung. Harnessing the power of synthetic data in healthcare: innovation, application, and privacy.npj Digital Medicine, 6, 10 2023

    work page 2023

  10. [10]

    Generative ai mitigates representation bias using synthetic health data.medRxiv, 2024

    Nicolo Micheletti, Raffaele Marchesi, Nicholas I-Hsien Kuo, Sebastiano Barbieri, Giuseppe Jurman, and Venet Osmani. Generative ai mitigates representation bias using synthetic health data.medRxiv, 2024

    work page 2024

  11. [11]

    Generative ai mitigates representation bias using synthetic health data, 09 2023

    Nicolo Micheletti, Raffaele Marchesi, Nicholas Kuo, Sebastiano Barbieri, Giuseppe Jurman, and Venet Osmani. Generative ai mitigates representation bias using synthetic health data, 09 2023

    work page 2023

  12. [12]

    A methodology for controlling bias and fairness in synthetic data generation.Applied Sciences, 12:4619, 05 2022

    Enrico Barbierato, Marco Della Vedova, Daniele Tessera, Daniele Toti, and Nicola Vanoli. A methodology for controlling bias and fairness in synthetic data generation.Applied Sciences, 12:4619, 05 2022

    work page 2022

  13. [13]

    Decaf: Generating fair synthetic data using causally-aware generative networks, 10 2021

    Boris Breugel, Trent Kyono, Jeroen Berrevoets, and Mihaela Schaar. Decaf: Generating fair synthetic data using causally-aware generative networks, 10 2021

    work page 2021

  14. [14]

    Rodríguez-Almeida, Himar Fabelo, Samuel Ortega, Alejandro Deniz, Francisco J

    Antonio J. Rodríguez-Almeida, Himar Fabelo, Samuel Ortega, Alejandro Deniz, Francisco J. Balea-Fernandez, Eduardo Quevedo, Cristina Soguero- Ruíz, Ana Maria Wägner, and Gustavo Marrero Callicó. Synthetic patient data generation and evaluation in disease prediction using small and imbalanced datasets.IEEE Journal of Biomedical and Health Informatics, 27:26...

    work page 2022

  15. [15]

    Measuring the quality of synthetic data for use in competitions, 06 2018

    James Jordon, Jinsung Yoon, and Mihaela Schaar. Measuring the quality of synthetic data for use in competitions, 06 2018

    work page 2018

  16. [16]

    How faithful is your synthetic data? sample-level metrics for evaluating and auditing generative models, 02 2021

    Ahmed Alaa, Boris van Breugel, Evgeny Saveliev, and Mihaela Schaar. How faithful is your synthetic data? sample-level metrics for evaluating and auditing generative models, 02 2021

    work page 2021

  17. [17]

    Synthetic data generation for tabular health records: A systematic review.Neurocomputing, 493, 04 2022

    Mikel Hernandez, Gorka Epelde, Ane Alberdi, Rodrigo Cilla, and Debbie Rankin. Synthetic data generation for tabular health records: A systematic review.Neurocomputing, 493, 04 2022. Manuscript submitted to ACM Critical Challenges and Guidelines in Evaluating Synthetic Tabular Data: A Systematic Review 15

    work page 2022

  18. [18]

    Gans trained by a two time-scale update rule converge to a local nash equilibrium

    Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bernhard Nessler, and Sepp Hochreiter. Gans trained by a two time-scale update rule converge to a local nash equilibrium. 12 2017

    work page 2017

  19. [19]

    Bertscore: Evaluating text generation with bert, 04 2019

    Tianyi Zhang, Varsha Kishore, Felix Wu, Kilian Weinberger, and Yoav Artzi. Bertscore: Evaluating text generation with bert, 04 2019

    work page 2019

  20. [20]

    Mimic-iii, a freely accessible critical care database.Scientific data, 3(1):1–9, 2016

    Alistair EW Johnson, Tom J Pollard, Lu Shen, Li-wei H Lehman, Mengling Feng, Mohammad Ghassemi, Benjamin Moody, Peter Szolovits, Leo Anthony Celi, and Roger G Mark. Mimic-iii, a freely accessible critical care database.Scientific data, 3(1):1–9, 2016

    work page 2016

  21. [21]

    Mimic-iv, a freely accessible electronic health record dataset.Scientific data, 10(1):1, 2023

    Alistair EW Johnson, Lucas Bulgarelli, Lu Shen, Alvin Gayles, Ayad Shammout, Steven Horng, Tom J Pollard, Sicheng Hao, Benjamin Moody, Brian Gow, et al. Mimic-iv, a freely accessible electronic health record dataset.Scientific data, 10(1):1, 2023

    work page 2023

  22. [22]

    The prisma 2020 statement: an updated guideline for reporting systematic reviews.Bmj, 372, 2021

    Matthew J Page, Joanne E McKenzie, Patrick M Bossuyt, Isabelle Boutron, Tammy C Hoffmann, Cynthia D Mulrow, Larissa Shamseer, Jennifer M Tetzlaff, Elie A Akl, Sue E Brennan, et al. The prisma 2020 statement: an updated guideline for reporting systematic reviews.Bmj, 372, 2021

    work page 2020

  23. [23]

    M. A. Jatoi and Nidal S. Kamel. Brain source localization using reduced eeg sensors.Signal, Image and Video Processing, 12:1447 – 1454, 2018

    work page 2018

  24. [24]

    Motion artifact synthesis for research in biomedical signal quality analysis.Biomedical Signal Processing and Control, 68:102611, 07 2021

    Emma Farago and Adrian Chan. Motion artifact synthesis for research in biomedical signal quality analysis.Biomedical Signal Processing and Control, 68:102611, 07 2021

    work page 2021

  25. [25]

    Chih-En Kuo, Tsung-Hua Lu, Guan-Ting Chen, and Po-Yu Liao. Towards precision sleep medicine: Self-attention gan as an innovative data augmentation technique for developing personalized automatic sleep scoring classification.Computers in Biology and Medicine, 148:105828, 2022

    work page 2022

  26. [26]

    Alireza Rafiei, Milad Ghiasi Rad, Andrea Sikora, and Rishikesan Kamaleswaran. Improving irregular temporal modeling by integrating synthetic data to the electronic medical record using conditional gans: a case study of fluid overload prediction in the intensive care unit.medRxiv, pages 2023–06, 2023

    work page 2023

  27. [27]

    Har-ctgan: a mobile sensor data generation tool for human activity recognition

    Joshua DeOliveira, Walter Gerych, Aruzhan Koshkarova, Elke Rundensteiner, and Emmanuel Agu. Har-ctgan: a mobile sensor data generation tool for human activity recognition. In2022 IEEE International Conference on Big Data (Big Data), pages 5233–5242. IEEE, 2022

    work page 2022

  28. [28]

    Twin: Personalized clinical trial digital twin generation

    Trisha Das, Zifeng Wang, and Jimeng Sun. Twin: Personalized clinical trial digital twin generation. InProceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages 402–413, 2023

    work page 2023

  29. [29]

    Conditional generation of medical time series for extrapolation to underrepresented populations.PLOS Digital Health, 1(7):e0000074, 2022

    Simon Bing, Andrea Dittadi, Stefan Bauer, and Patrick Schwab. Conditional generation of medical time series for extrapolation to underrepresented populations.PLOS Digital Health, 1(7):e0000074, 2022

    work page 2022

  30. [30]

    Quantifying resemblance of synthetic medical time-series

    Karan Bhanot, Saloni Dash, Joseph Pedersen, Isabelle Guyon, and Kristin P Bennett. Quantifying resemblance of synthetic medical time-series. In ESANN, 2021

    work page 2021

  31. [31]

    Evaluating identity disclosure risk in fully synthetic health data: model development and validation.Journal of medical Internet research, 22(11):e23139, 2020

    Khaled El Emam, Lucy Mosquera, and Jason Bass. Evaluating identity disclosure risk in fully synthetic health data: model development and validation.Journal of medical Internet research, 22(11):e23139, 2020

    work page 2020

  32. [32]

    Ya-Ting Lu, Horng-Jiun Chao, Yi-Chun Chiang, and Hsiang-Yin Chen. Explainable machine learning techniques to predict amiodarone-induced thyroid dysfunction risk: multicenter, retrospective study with external validation.Journal of Medical Internet Research, 25:e43734, 2023

    work page 2023

  33. [33]

    Modeling heart rate variability including the effect of sleep stages.Chaos: An Interdisciplinary Journal of Nonlinear Science, 26(2), 2016

    Mateusz Soliński, Jan Gierałtowski, and Jan Żebrowski. Modeling heart rate variability including the effect of sleep stages.Chaos: An Interdisciplinary Journal of Nonlinear Science, 26(2), 2016

    work page 2016

  34. [34]

    The problem of fairness in synthetic healthcare data.Entropy, 23(9):1165, 2021

    Karan Bhanot, Miao Qi, John S Erickson, Isabelle Guyon, and Kristin P Bennett. The problem of fairness in synthetic healthcare data.Entropy, 23(9):1165, 2021

    work page 2021

  35. [35]

    Publishing differentially private medical events data

    Sigal Shaked and Lior Rokach. Publishing differentially private medical events data. InA vailability, Reliability, and Security in Information Systems: IFIP WG 8.4, 8.9, TC 5 International Cross-Domain Conference, CD-ARES 2016, and Workshop on Privacy A ware Machine Learning for Health Data Science, PAML 2016, Salzburg, Austria, August 31-September 2, 201...

    work page 2016

  36. [36]

    Morphology-preserving reconstruction of times series with missing data for enhancing deep learning-based classification.Biomedical Signal Processing and Control, 70:103052, 2021

    Nooshin Bahador, Guoying Zhao, Jarno Jokelainen, Seppo Mustola, and Jukka Kortelainen. Morphology-preserving reconstruction of times series with missing data for enhancing deep learning-based classification.Biomedical Signal Processing and Control, 70:103052, 2021

    work page 2021

  37. [37]

    A joint learning method for incomplete and imbalanced data in electronic health record based on generative adversarial networks.Computers in Biology and Medicine, 168:107687, 2024

    Xutao Weng, Hong Song, Yucong Lin, You Wu, Xi Zhang, Bowen Liu, and Jian Yang. A joint learning method for incomplete and imbalanced data in electronic health record based on generative adversarial networks.Computers in Biology and Medicine, 168:107687, 2024

    work page 2024

  38. [38]

    Investigating synthetic medical time-series resemblance.Neurocomputing, 494:368–378, 2022

    Karan Bhanot, Joseph Pedersen, Isabelle Guyon, and Kristin P Bennett. Investigating synthetic medical time-series resemblance.Neurocomputing, 494:368–378, 2022

    work page 2022

  39. [39]

    Synthetic electronic health records generated with variational graph autoencoders.NPJ Digital Medicine, 6(1):83, 2023

    Giannis Nikolentzos, Michalis Vazirgiannis, Christos Xypolopoulos, Markus Lingman, and Erik G Brandt. Synthetic electronic health records generated with variational graph autoencoders.NPJ Digital Medicine, 6(1):83, 2023

    work page 2023

  40. [40]

    Generation of surrogate event sequences via joint distribution of successive inter-event intervals.Chaos: An Interdisciplinary Journal of Nonlinear Science, 29(12), 2019

    Leonardo Ricci, Michele Castelluzzo, Ludovico Minati, and Alessio Perinelli. Generation of surrogate event sequences via joint distribution of successive inter-event intervals.Chaos: An Interdisciplinary Journal of Nonlinear Science, 29(12), 2019

    work page 2019

  41. [41]

    A new attack method against ecg-based key generation and agreement schemes in body area networks.IEEE Sensors Journal, 21(15):17300–17307, 2021

    Jack Hodgkiss, Soufiene Djahel, and Zonghua Zhang. A new attack method against ecg-based key generation and agreement schemes in body area networks.IEEE Sensors Journal, 21(15):17300–17307, 2021

    work page 2021

  42. [42]

    Epileptic seizure prediction based on eeg by auto-machine learning

    Cai Chen, Fulai Peng, Yue Sun, Danyang Lv, Ningling Zhang, Xingwei Wang, and Lin Wang. Epileptic seizure prediction based on eeg by auto-machine learning. In2022 IEEE International Conference on Real-time Computing and Robotics (RCAR), pages 710–715. IEEE, 2022

    work page 2022

  43. [43]

    Patient flow prediction via discriminative learning of mutually-correcting processes.IEEE transactions on Knowledge and Data Engineering, 29(1):157–171, 2016

    Hongteng Xu, Weichang Wu, Shamim Nemati, and Hongyuan Zha. Patient flow prediction via discriminative learning of mutually-correcting processes.IEEE transactions on Knowledge and Data Engineering, 29(1):157–171, 2016

    work page 2016

  44. [44]

    Ameer Mohammed, Majid Zamani, Richard Bayford, and Andreas Demosthenous. Toward on-demand deep brain stimulation using online parkinson’s disease prediction driven by dynamic detection.IEEE Transactions on Neural Systems and Rehabilitation Engineering, 25(12):2441–2452, 2017. Manuscript submitted to ACM 16 Nafis et al

    work page 2017

  45. [45]

    Afe-gan: Synthesizing electrocardiograms with atrial fibrillation characteristics using generative adversarial networks

    Xianglong Wang, Berkman Sahiner, Christopher G Scully, and Kenny H Cha. Afe-gan: Synthesizing electrocardiograms with atrial fibrillation characteristics using generative adversarial networks. In2023 45th Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC), pages 1–5. IEEE, 2023

    work page 2023

  46. [46]

    Synthetic and private smart health care data generation using gans

    Sana Imtiaz, Muhammad Arsalan, Vladimir Vlassov, and Ramin Sadre. Synthetic and private smart health care data generation using gans. In2021 International Conference on Computer Communications and Networks (ICCCN), pages 1–7. IEEE, 2021

    work page 2021

  47. [47]

    Protect and extend-using gans for synthetic data generation of time-series medical records

    Navid Ashrafi, Vera Schmitt, Robert P Spang, Sebastian Möller, and Jan-Niklas Voigt-Antons. Protect and extend-using gans for synthetic data generation of time-series medical records. In2023 15th International Conference on Quality of Multimedia Experience (QoMEX), pages 171–176. IEEE, 2023

    work page 2023

  48. [48]

    Alireza Rafiei, Milad Ghiasi Rad, Andrea Sikora, and Rishikesan Kamaleswaran. Improving mixed-integer temporal modeling by generating synthetic data using conditional generative adversarial networks: A case study of fluid overload prediction in the intensive care unit.Computers in Biology and Medicine, 168:107749, 2024

    work page 2024

  49. [49]

    Analysis of temporal correlation in heart rate variability through maximum entropy principle in a minimal pairwise glassy model.Scientific Reports, 10(1):15353, 2020

    Elena Agliari, Francesco Alemanno, Adriano Barra, Orazio Antonio Barra, Alberto Fachechi, Lorenzo Franceschi Vento, and Luciano Moretti. Analysis of temporal correlation in heart rate variability through maximum entropy principle in a minimal pairwise glassy model.Scientific Reports, 10(1):15353, 2020

    work page 2020

  50. [50]

    Generating synthetic mixed-type longitudinal electronic health records for artificial intelligent applications.NPJ Digital Medicine, 6(1):98, 2023

    Jin Li, Benjamin J Cairns, Jingsong Li, and Tingting Zhu. Generating synthetic mixed-type longitudinal electronic health records for artificial intelligent applications.NPJ Digital Medicine, 6(1):98, 2023

    work page 2023

  51. [51]

    Ziqi Zhang, Chao Yan, and Bradley A Malin. Keeping synthetic patients on track: feedback mechanisms to mitigate performance drift in longitudinal health data simulation.Journal of the American Medical Informatics Association, 29(11):1890–1898, 2022

    work page 2022

  52. [52]

    In-Sun Oh, Han Eol Jeong, Hyesung Lee, Kristian B Filion, Yunha Noh, and Ju-Young Shin. Validating an approach to overcome the immeasurable time bias in cohort studies: a real-world example and monte carlo simulation study.International Journal of Epidemiology, 52(5):1534–1544, 2023

    work page 2023

  53. [53]

    Improving ppg-based heart-rate monitoring with synthetically generated data

    Alessio Burrello, Daniele Jahier Pagliari, Marzia Bianco, Enrico Macii, Luca Benini, Massimo Poncino, and Simone Benatti. Improving ppg-based heart-rate monitoring with synthetically generated data. In2022 IEEE Biomedical Circuits and Systems Conference (BioCAS), pages 153–157. IEEE, 2022

    work page 2022

  54. [54]

    Emg data augmentation for grasp classification using generative adversarial networks

    Vincent Mendez, Clément Lhoste, and Silvestro Micera. Emg data augmentation for grasp classification using generative adversarial networks. In 2022 44th Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC), pages 3619–3622. IEEE, 2022

    work page 2022

  55. [55]

    Discovery of early-alert indicators using hybrid ensemble learning and generative physics-based models

    Zhenyi Yang, Rebecca Miao, Marina Orlova, Ivan Nechepurenko, and Valeriy Gavrishchaka. Discovery of early-alert indicators using hybrid ensemble learning and generative physics-based models. In2022 5th International Conference on Information and Computer Technologies (ICICT), pages 224–232. IEEE, 2022

    work page 2022

  56. [56]

    Gan based synthetic ecg for psychological stress

    Adrian Vulpe-Grigoraşi and Ovidiu Grigore. Gan based synthetic ecg for psychological stress. In2022 E-Health and Bioengineering Conference (EHB), pages 1–4. IEEE, 2022

    work page 2022

  57. [57]

    Virtual patient model: an approach for generating synthetic healthcare time series data

    Rittika Shamsuddin, Barbara M Maweu, Ming Li, and Balakrishnan Prabhakaran. Virtual patient model: an approach for generating synthetic healthcare time series data. In2018 IEEE International Conference on Healthcare Informatics (ICHI), pages 208–218. IEEE, 2018

    work page 2018

  58. [58]

    Discovering multidimensional motifs in physiological signals for personalized healthcare.IEEE journal of selected topics in signal processing, 10(5):832–841, 2016

    Arvind Balasubramanian, Jun Wang, and Balakrishnan Prabhakaran. Discovering multidimensional motifs in physiological signals for personalized healthcare.IEEE journal of selected topics in signal processing, 10(5):832–841, 2016

    work page 2016

  59. [59]

    Pns-gan: Conditional generation of peripheral nerve signals in the wavelet domain via adversarial networks

    Olivier Tessier-Larivière, Luke Y Prince, Pascal Fortier-Poisson, Lorenz Wernisch, Oliver Armitage, Emil Hewage, Guillaume Lajoie, and Blake A Richards. Pns-gan: Conditional generation of peripheral nerve signals in the wavelet domain via adversarial networks. In2021 10th International IEEE/EMBS Conference on Neural Engineering (NER), pages 778–782. IEEE, 2021

    work page 2021

  60. [60]

    A framework for patient state tracking by classifying multiscalar physiologic waveform features.IEEE Transactions on Biomedical Engineering, 64(12):2890–2900, 2017

    Benjamin Vandendriessche, Mustafa Abas, Thomas E Dick, Kenneth A Loparo, and Frank J Jacono. A framework for patient state tracking by classifying multiscalar physiologic waveform features.IEEE Transactions on Biomedical Engineering, 64(12):2890–2900, 2017

    work page 2017

  61. [61]

    An approach for analyzing cognitive behavior of autism spectrum disorder using p300 bci data

    Nabila Tasnim, Joyita Halder, Shahed Ahmed, and Shaikh Anowarul Fattah. An approach for analyzing cognitive behavior of autism spectrum disorder using p300 bci data. In2022 IEEE Region 10 Symposium (TENSYMP), pages 1–6. IEEE, 2022

    work page 2022

  62. [62]

    enhancing small tabular clinical trial dataset through hybrid data augmentation: combining smote and wcgan-gp

    Winston Wang and Tun-Wen Pai. enhancing small tabular clinical trial dataset through hybrid data augmentation: combining smote and wcgan-gp. Data, 8(9):135, 2023

    work page 2023

  63. [63]

    Clara García-Vicente, David Chushig-Muzo, Inmaculada Mora-Jiménez, Himar Fabelo, Inger Torhild Gram, Maja-Lisa Løchen, Conceição Granja, and Cristina Soguero-Ruiz. Evaluation of synthetic categorical data generation techniques for predicting cardiovascular diseases and post-hoc interpretability of the risk factors.Applied Sciences, 13(7):4119, 2023

    work page 2023

  64. [64]

    The health gym: synthetic health-related datasets for the development of reinforcement learning algorithms

    Nicholas I-Hsien Kuo, Mark N Polizzotto, Simon Finfer, Federico Garcia, Anders Sönnerborg, Maurizio Zazzi, Michael Böhm, Rolf Kaiser, Louisa Jorm, and Sebastiano Barbieri. The health gym: synthetic health-related datasets for the development of reinforcement learning algorithms. Scientific data, 9(1):693, 2022

    work page 2022

  65. [65]

    Ehr-safe: generating high-fidelity and privacy-preserving synthetic electronic health records.NPJ Digital Medicine, 6(1):141, 2023

    Jinsung Yoon, Michel Mizrahi, Nahid Farhady Ghalaty, Thomas Jarvinen, Ashwin S Ravi, Peter Brune, Fanyu Kong, Dave Anderson, George Lee, Arie Meir, et al. Ehr-safe: generating high-fidelity and privacy-preserving synthetic electronic health records.NPJ Digital Medicine, 6(1):141, 2023

    work page 2023

  66. [66]

    Cardiosim: a pc-based cardiac signal simulator using segmental modeling of electrocardiogram.Computer Methods in Biomechanics and Biomedical Engineering, 26(13):1532–1548, 2023

    Anumita Mitra, Palash Kumar Kundu, Rajarshi Gupta, Jayanta Saha, and Arunansu Talukdar. Cardiosim: a pc-based cardiac signal simulator using segmental modeling of electrocardiogram.Computer Methods in Biomechanics and Biomedical Engineering, 26(13):1532–1548, 2023

    work page 2023

  67. [67]

    Using genetic algorithms for optimal change point detection in activity monitoring

    Naveed Khan, Sally McClean, Shuai Zhang, and Chris Nugent. Using genetic algorithms for optimal change point detection in activity monitoring. In2016 IEEE 29th International Symposium on Computer-Based Medical Systems (CBMS), pages 318–323. IEEE, 2016. Manuscript submitted to ACM Critical Challenges and Guidelines in Evaluating Synthetic Tabular Data: A S...

    work page 2016

  68. [68]

    A comparison of three methodologies for the computation of v-index

    Ebadollah Kheirati Roonizi, Massimo W Rivolta, Luca T Mainardi, and Roberto Sassi. A comparison of three methodologies for the computation of v-index. In2015 Computing in Cardiology Conference (CinC), pages 593–596. IEEE, 2015

    work page 2015

  69. [69]

    Overcoming data scarcity in human activity recognition

    Orhan Konak, Lucas Liebe, Kirill Postnov, Franz Sauerwald, Hristijan Gjoreski, Mitja Luštrek, and Bert Arnrich. Overcoming data scarcity in human activity recognition. In2023 45th Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC), pages 1–7. IEEE, 2023

    work page 2023

  70. [70]

    Tabular gan-based oversampling of imbalanced time-to-event data for survival prediction

    Huaning Tan, Renxing Chen, Meng Qin, Lining Tang, Zhibing Wu, Qianlin Luo, and Yujuan Quan. Tabular gan-based oversampling of imbalanced time-to-event data for survival prediction. In2023 8th International Conference on Cloud Computing and Big Data Analytics (ICCCBDA), pages 376–380. IEEE, 2023

    work page 2023

  71. [71]

    Estimating individual treatment effects with time-varying confounders

    Ruoqi Liu, Changchang Yin, and Ping Zhang. Estimating individual treatment effects with time-varying confounders. In2020 IEEE International Conference on Data Mining (ICDM), pages 382–391. IEEE, 2020

    work page 2020

  72. [72]

    Generated tabular data with multi-gans for arrhythmia classification based on dnn models

    Hendrico Yehezky Nata Atmadja, Alhadi Bustamam, et al. Generated tabular data with multi-gans for arrhythmia classification based on dnn models. In2022 International Conference of Science and Information Technology in Smart Administration (ICSINTESA), pages 69–74. IEEE, 2022

    work page 2022

  73. [73]

    Time-series anonymization of tabular health data using generative adversarial network

    Atiye Sadat Hashemi, Kobra Etminani, Amira Soliman, Omar Hamed, and Jens Lundström. Time-series anonymization of tabular health data using generative adversarial network. In2023 International Joint Conference on Neural Networks (IJCNN), pages 1–8. IEEE, 2023

    work page 2023

  74. [74]

    Plasek, Yun Xiong, Yangyong Zhu, D

    Xiaoxia Wang, Chun-Pong Tang, Yanming He, Joseph M. Plasek, Yun Xiong, Yangyong Zhu, D. Bates, and Li Zhou. Using an optimized generative model to infer the progression of complications in type 2 diabetes patients.BMC Medical Informatics and Decision Making, 22, 2022

    work page 2022

  75. [75]

    Ecg quality assessment via deep learning and data augmentation

    Álvaro Huerta, Arturo Martínez-Rodrigo, José Joaquín Rieta, and Raúl Alcaraz. Ecg quality assessment via deep learning and data augmentation. 2021 Computing in Cardiology (CinC), 48:1–4, 2021

    work page 2021

  76. [76]

    Ayilara, Robert W

    Olawale F. Ayilara, Robert W. Platt, Matt Dahl, Janie Coulombe, Pablo Gonzalez Ginestet, Dan Chateau, and Lisa M. Lix. Generating synthetic data from administrative health records for drug safety and effectiveness studies.International Journal of Population Data Science, 8, 2023

    work page 2023

  77. [77]

    Deep-learning-driven techniques for real-time multimodal health and physical data synthesis.Electronics, 2023

    Muhammad Salman Haleem, Audrey Ekuban, Alessio Antonini, Silvio Marcello Pagliara, Leandro Pecchia, and Carlo Allocca. Deep-learning-driven techniques for real-time multimodal health and physical data synthesis.Electronics, 2023

    work page 2023

  78. [78]

    Ecg synthesis via diffusion-based state space augmented transformer.Sensors (Basel, Switzerland), 23, 2023

    Md Haider Zama and Friedhelm Schwenker. Ecg synthesis via diffusion-based state space augmented transformer.Sensors (Basel, Switzerland), 23, 2023

    work page 2023

  79. [79]

    Lucy Mosquera, Khaled El Emam, Lei Ding, Vishal Sharma, Xue Hua Zhang, Samer El Kababji, Chris Carvalho, Brian Hamilton, Dan Palfrey, Linglong Kong, Bei Jiang, and Dean T. Eurich. A method for generating synthetic longitudinal health data.BMC Medical Research Methodology, 23, 2023

    work page 2023

  80. [80]

    Shah, and Lixia Yao

    Ming Huang, Nilay D. Shah, and Lixia Yao. Evaluating global and local sequence alignment methods for comparing patient medical records.BMC Medical Informatics and Decision Making, 19, 2019

    work page 2019

Showing first 80 references.