Pith. sign in

REVIEW 4 major objections 5 minor 52 references

Identifying Heart Attack Risk in Vulnerable Population: A Machine Learning Approach

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A hybrid clustering and regression model identifies postmenopausal women as the highest-risk heart attack group, and the paper claims this menopause dependence is a new finding.

desk verdict The central risk-scoring method is invalid — it substitutes inter-feature correlations for regression coefficients — and the menopause finding is an artifact of that score, so this paper should not go forward. read the letter →

arxiv 2505.21139 v1 pith:O7BBQYSY submitted 2025-05-27 q-bio.PE cond-mat.softcs.LGphysics.med-ph

Subhagata Chattopadhyay , Amit K Chattopadhyay This is my paper
classification q-bio.PEcond-mat.softcs.LGphysics.med-ph
keywords heartattackriskmachinelearningGaussianmixturemodelclusteringPearsoncorrelationlinearregressionscoremenopauseclinicaldecisionsupportsystem
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper attempts to establish that a hybrid machine-learning workflow can turn a public 303-patient heart-attack dataset into per-patient risk scores and a ranked vulnerable population. It reports that Gaussian Mixture Model clustering outperforms four alternatives in isolating at-risk cases, and that using Pearson correlation values as coefficients in a linear regression yields scores whose top cases are postmenopausal women aged 62-67 with very high cholesterol and elevated thallium stress-test results. The authors claim the precise dependence of heart attack risk on menopause is a new finding with direct implications for clinical decision support. If the workflow is valid, it would give clinicians a non-invasive way to prioritize middle-aged and elderly women for early intervention.

What carries the argument

The mechanism that carries the argument is Equation (2), $y = \sum_i x_i b_i + c$, a linear regression for heart attack risk score in which the regression coefficients $b_i$ are replaced by Pearson correlation values $r$ computed between predictor pairs inside the GMM-selected high-risk cluster (Table 4), with the intercept $c$ set to zero. Gaussian Mixture Model clustering supplies the at-risk subpopulation; Pearson's correlation supplies the descriptor set; the substituted-$r$ regression converts each case's 13 attribute values into a single score. The named object is the hybrid 'clustering plus linear regression' model, and its output risk scores rank cases so that the top three define the vulnerable subpopulation featured in the discussion.

What would settle it

A reader could settle the claim by regressing the actual heart-attack indicator (AR/NAR) on the same 13 predictors with ordinary least squares, computing per-case scores from those fitted coefficients, and comparing the two rankings; the r-based scores are doing the work the paper claims only if they separate the AR and NAR groups at least as well as the fitted model. A simpler version: compute the area under the ROC curve for the published risk score against the outcome label; a value near 0.5 would show the score carries no discriminative information about heart attack status.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that a Gaussian Mixture Model applied to the 303-case dataset separates a high-risk cluster with 84.24% accuracy, and that the twelve inter-feature Pearson coefficients computed inside that cluster, when substituted for the coefficients in a multiple linear regression with zero intercept, produce a heart attack risk score. The top-scoring cases are all women with mean age 64.66 years, mean cholesterol 458.33 mg%, high maximum heart rate, and elevated thallium readings, with no ECG changes, no exercise angina, and no visible coronary blockage. From this the paper concludes that postmenopausal estrogen depletion, possibly compounded by anxiety and pandemic-era stress, is the central aggravating factor, and that menopause is a critical physiological threshold for heart attack risk whose precise dependence has not previously been quantified. The paper also presents the workflow as a deployable clinical decision support system for risk stratification.

Load-bearing premise

The load-bearing premise is that Pearson correlations computed between feature pairs inside the high-risk cluster can validly stand in for regression coefficients in a heart attack risk score, even though those correlations measure association between two measurements rather than association with heart attack status, and the resulting scores are never checked against whether a patient actually had a heart attack.

Editorial extensions

If this is right

  • Clinicians could use the computed risk scores to prioritize at-risk patients in a clinical decision support system, treating the top scorers as needing earliest review.
  • Screening and prevention efforts would focus on postmenopausal and perimenopausal women, a group the paper argues deserves the same vigilance as men.
  • The model could be repurposed, with additional extrinsic stress factors such as anxiety, financial hardship, and pandemic-related stressors, to assess post-COVID cardiovascular risk.
  • The GMM-based clustering comparison offers a practical baseline for choosing a clustering method in cardiovascular risk stratification, with accuracy prioritized over silhouette-based separation.

Reading between the lines

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

  • The score's lack of validation against outcome labels suggests a direct next step: check whether the computed score separates AR from NAR cases in a held-out set; the paper itself lists real-world validation by clinicians as future work.
  • The correlation-coefficient substitution could be tested by comparing the r-based score ranking with a ranking from ordinary least-squares coefficients fit to the same 13 predictors on the outcome.
  • Adding menopause as an explicit predictor, rather than inferring it from age and sex, would make the hypothesized estrogen-depletion mechanism directly measurable.
  • The top-three profile (high cholesterol, high thallium, normal ECG, no vessel involvement) suggests a vasospasm-predominant postmenopausal phenotype that a prospective biomarker or imaging study could investigate.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The manuscript proposes a hybrid machine-learning pipeline for heart-attack risk scoring on the Kaggle heart-attack dataset (303×13). Five clustering algorithms are compared after removing the outcome label; the Gaussian Mixture Model (GMM) is selected as best (84.24% AR accuracy, Silhouette 0.2623). Within the GMM-defined high-risk cluster, Pearson correlations between predictor pairs are computed, and Section 3 stipulates that these r-values replace the regression coefficients b_i in Equation (2) (with intercept c=0) to produce a heart-attack risk score for each case. The top three scoring cases are postmenopausal women, and the Discussion interprets this as a new finding on the precise dependence of heart attacks on menopause, with implications for a clinical decision support system and for COVID-19-era risk. The paper states explicitly that no direct COVID-19 link is established and lists several limitations, including small sample size and lack of real-world validation.

Significance. If the method were valid, the paper would offer a simple, interpretable risk score and a concrete subpopulation (postmenopausal women) for targeted intervention, which would be clinically useful. The manuscript also has some strengths: it uses a public, reproducible dataset; it compares five clustering algorithms with transparent accuracy and Silhouette metrics; it reports descriptive statistics and internal-consistency checks; and it candidly lists limitations, including the lack of a direct COVID-19 link. However, these strengths do not compensate for the core methodological flaw described below. Because the central risk-scoring step has no demonstrated relation to the outcome and is never validated against heart-attack status, the paper's main claims—reliable risk scoring and a new menopause-dependence finding—are unsupported.

major comments (4)
  1. [Section 3, Equation (2) and Table 4] The load-bearing step of the paper is invalid as stated. Table 4 lists Pearson correlations between pairs of predictors (e.g., Age–trtbps, cp–thalachh, thalachh–slp), not correlations between predictors and the heart-attack outcome. The text then says these r-values replace the regression coefficients b_i in Equation (2). A Pearson correlation between two predictors is not an outcome-based regression weight: it measures mutual association among features, not association with the target. There is no mathematical guarantee that the resulting score is monotonically related to the probability of heart attack, and the manuscript provides no empirical check of such a relation. The scores in Tables 5 and 6 are therefore not demonstrated risk scores; they are weighted feature sums whose clinical interpretation is unfounded.
  2. [Section 3, GMM cluster selection and post hoc removal of misclassified cases] The evaluation is circular. The GMM cluster is chosen to maximize agreement with the outcome labels (84.24% AR accuracy), and the 15.76% of AR cases that GMM misclassifies are 'removed from further study.' The remaining 139 cases are then used to compute the very correlations that define the risk score, and those same cases are scored and ranked. No independent test set is used for the scoring step. The 5-fold cross-validation statement in Section 3 refers to clustering variation (average variation 0.0015 in Section 4, Table 7 discussion), not to validation of the risk score against held-out outcomes. Thus the top-three finding and the menopause claim derive from a pipeline that has never been tested on data not used to construct the score.
  3. [Section 4 and Conclusions, menopause claim] The central novel claim—'the precise dependence of heart attacks on menopause is an altogether new finding from this study'—is not supported by the data or the analysis. The dataset contains no menopausal status or hormonal measurements; menopause is inferred solely from age and sex for the top three scoring cases. Even if the risk score were valid, three cases cannot establish a precise dependence. The Discussion itself acknowledges the inference is physiological ('supposedly postmenopausal'), and the paper states that the association is assumed rather than directly measured. The conclusion therefore overstates what the analysis can establish.
  4. [Table 4 and Section 3, p-values] The claim that p-values are '< 0.05' for all listed correlations is unverifiable and, as reported, inconsistent with the table structure: the table provides only two p-value columns, and the text does not state how multiple comparisons were handled. More importantly, the statistical significance of inter-feature correlations does not validate them as outcome-predictive coefficients. The manuscript appears to conflate 'statistically significant correlation between predictors' with 'significant predictor of heart attack risk,' which is a logical error regardless of the p-value values.
minor comments (5)
  1. [Abstract and Introduction] The abstract claims 'strong association between the likelihood of experiencing a heart attack on the 13 risk factors studied,' but the analysis only computes correlations among predictors, not between predictors and the outcome. The wording should be corrected to describe what was actually computed.
  2. [Section 2, dataset description] The dataset is called '303 × 13' twice but has 13 predictors plus a target, so the shape should be 303 × 14 (or 303 × 13 features plus one target). This inconsistency is confusing.
  3. [Section 3, Table 4 numbering] The table lists row numbers 1, 2, 3, 4, 5, 6, 8, 9, 10, 11, 12; row 7 is missing. Either the numbering is erroneous or a row was omitted. Also, the text in Section 3 states the maximum correlation is 0.38 while Table 4 gives 0.39 for thalachh–slp; the numbers should be reconciled.
  4. [Section 3, Equation (1)] The formula for Pearson's r is garbled: the denominator and the summation notation are not typeset correctly, making the equation hard to read. Equation (4) similarly has a formatting error in the denominator.
  5. [References] Several references are incomplete or inconsistent: reference [1] lacks a year; [5] duplicates [4]; [16] has an unbalanced parenthesis; [24] misspells 'Structural.' These should be corrected during revision.

Circularity Check

2 steps flagged · score 8.0 of 10

The heart attack risk score is defined, not derived: Equation (2) with b replaced by inter-feature Pearson r values makes 'high risk' synonymous with 'high weighted sum of features,' so the menopause conclusion restates the score construction.

  1. self definitional [Section 3, Equation (2) and Table 4 (Results of Pearson's correlation on the GMM-based high-risk cluster)]
    "These high 'r' values are then leveraged to form a linear regression equation [22] for heart attack risk scoring (Equation (2)) ... 'b' values are replaced by high 'r' values in this work to compute the respective 'y'."

    The risk score y is defined as a weighted sum y = Σ x_i r_i, where the r_i are Pearson correlations between pairs of predictor variables within the GMM-selected high-risk cluster, not correlations with the heart attack outcome. The same score is then used to rank cases as 'high-risk' and to support the conclusion that postmenopausal women are at highest risk. Because 'high risk' is defined by this score and the score is defined from the same cluster's feature correlations, the menopause finding is a restatement of the score's construction rather than an independent empirical result.

  2. fitted input called prediction [Section 3, Tables 5 and 6, and the 5-fold cross-validation note]
    "Individual heart attack risk score is computed using Equation (2) and showcased in Table 6. ... We note that the average variation in the risk scores using 5-fold cross validation is 0.0015, which is clinically negligible."

    The r coefficients are computed from the GMM high-risk cluster, and then the same cases in that cluster are 'scored' and 'predicted' using those coefficients. This is an in-sample rescoring, not an out-of-sample prediction. The 5-fold cross-validation variation of 0.0015 is reported without any split details or validation against outcome labels, so it cannot establish that the score predicts heart attack status. The top three 'predicted at-risk' cases are simply the cases with the largest weighted sums under a formula whose weights came from the very same data.

full rationale

The paper's derivation chain is: (1) cluster the 303 Kaggle cases with outcome labels removed but evaluate clustering accuracy against those labels; (2) select the GMM high-risk cluster; (3) compute Pearson correlations between predictor pairs within that cluster; (4) substitute those r values for regression coefficients in Equation (2) to compute a 'heart attack risk score'; (5) rank the same cases by this score; and (6) infer from the top three cases that menopause in women is a critical threshold aggravating heart attack risk. The circularity is concentrated in steps (4)–(6): Equation (2) with b replaced by r defines the score, and the conclusion that the top-scoring cases are the highest-risk cases is true by definition. No regression against the outcome label is ever performed, and the r values in Table 4 quantify associations among predictors (e.g., thalachh–slp, cp–thalachh), not associations with heart attacks. Thus the 'precise dependence of heart attacks on menopause' is an artifact of which features receive positive weights and which cases have high values of those features, rather than a measured association with myocardial infarction. The 5-fold cross-validation statement is unreferenced and provides no independent validation. The paper itself admits in Limitations that 'Risk validation by the medical doctors must involve data from real-world cardiac risk datasets,' confirming the absence of external validation. Because the central claim reduces by construction to the arbitrary scoring formula, the circularity score is high; however, there is no evidence of a load-bearing self-citation chain, so the score is 8 rather than 10.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central claim rests on no new entities, but on a problematic set of fitted coefficients and domain assumptions. The r-as-b substitution is the most significant free parameter choice. The COVID-19 assumption and inferred menopausal status are unvalidated domain assumptions.

free parameters (2)
  • Regression coefficients b_i set equal to Pearson r values = 12 values from Table 4: 0.28, 0.21, 0.30, 0.19, 0.21, 0.14, 0.39, 0.29, 0.22, 0.27, 0.21
    In Equation (2), the weights b_i are replaced by correlation coefficients computed from the same GMM-selected high-risk cluster, without fitting to the outcome variable.
  • Constant term c in risk score equation = 0
    The constant is arbitrarily set to zero, forcing the regression line through the origin; this is a modeling choice not derived from data.
assumptions (4)
  • ad hoc to paper The Kaggle dataset is assumed to reflect COVID-19-era heart attack risk despite no direct link being established.
    Stated in Section 2: 'the influence of COVID-19 with heart health risk has been assumed in this work despite no direct link between COVID-19 pandemic and/or COVID-19 vaccination with the heart health dataset has been found.'
  • domain assumption Menopause is inferred from age and sex of the top three risk cases rather than measured.
    Section 4 discusses postmenopausal status based on the top three cases (females aged 62-67) even though the dataset contains no menopause variable.
  • domain assumption Pearson correlation and linear regression are applied to non-normal data with low internal consistency.
    Section 2 reports Cronbach alpha 0.08 and Shapiro-Wilk p=0, yet proceeds with Pearson correlation and least squares style regression.
  • domain assumption A Silhouette score above 0.25 is considered acceptable for cluster validity.
    Section 4 accepts the GMM Silhouette score of 0.2623 as reasonable based on reference [25], a domain-specific threshold.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Identifying Heart Attack Risk in Vulnerable Population: A Machine Learning Approach." pith.science (2026). https://pith.science/paper/O7BBQYSY

@misc{pith2026250521139,
  author       = {Pith},
  title        = {Pith review of: Identifying Heart Attack Risk in Vulnerable Population: A Machine Learning Approach},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/O7BBQYSY}},
  note         = {Machine review of arXiv:2505.21139}
}
read the original abstract

The COVID-19 pandemic has significantly increased the incidence of post-infection cardiovascular events, particularly myocardial infarction, in individuals over 40. While the underlying mechanisms remain elusive, this study employs a hybrid machine learning approach to analyze epidemiological data in assessing 13 key heart attack risk factors and their susceptibility. Based on a unique dataset that combines demographic, biochemical, ECG, and thallium stress-tests, this study categorizes distinct subpopulations against varying risk profiles and then divides the population into 'at-risk' (AR) and 'not-at-risk' (NAR) groups using clustering algorithms. The study reveals strong association between the likelihood of experiencing a heart attack on the 13 risk factors studied. The aggravated risk for postmenopausal patients indicates compromised individual risk factors due to estrogen depletion that may be, further compromised by extraneous stress impacts, like anxiety and fear, aspects that have traditionally eluded data modeling predictions.

Figures

Figures reproduced from arXiv: 2505.21139 by the authors.

Figure 5
Figure 5. Pearson’s correlation coefficient heatmap Since the ‘r’ values are not sufficiently high for independent validation, these are tested for statistical significance. The p-values fall much below 0.05 and hence the linearity among the variables is statistically significant. A detailed hand calculation is provided below for Case 1, while the heart attack risk scores for the top three high￾risk cases are summarized in [… view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

52 extracted references · 49 canonical work pages

  1. [26]

    Postmenopausal Women at Higher Risk of Heart Attacks Than Men Their Age,

    "Postmenopausal Women at Higher Risk of Heart Attacks Than Men Their Age," Inside Precision Med- icine, 11 May 2023. [Online]. Available: https://www.insideprecisionmedicine.com/news-and-features/post- menopausal-women-at-higher-risk-of-heart-attacks-than-men-their-age/. [Accessed 31 January 2024]

    work page 2023

  2. [1]

    Cardiovascular diseases,

    "Cardiovascular diseases," World Health Organization, 26 June 2023. [Online]. Available: https://www.who.int/health-topics/cardiovascular-diseases. [Accessed 19 January 2024]

    work page 2023

  3. [2]

    Prevalence of probable post-COVID cardiac sequelae and its health seeking behaviour among health care workers: A cross-sectional analytical study.,

    Y. Marimuthu, G. C. Nair, U. Nagesh, A. Anand, K. K. Chopra, B. Nagappa, N. Sharma, G. Sivashankar and N. Nagaraj, "Prevalence of probable post-COVID cardiac sequelae and its health seeking behaviour among health care workers: A cross-sectional analytical study.," Indian J Tuberc., p. (online ahead of print), 2023

    work page 2023

  4. [3]

    Long-term effect of SARS-CoV-2 infection on cardiovas- cular outcomes and all-cause mortality,

    A. Aleksova, A. L. Fluca, G. Gagno and e. al., "Long-term effect of SARS-CoV-2 infection on cardiovas- cular outcomes and all-cause mortality," Life Sci., vol. 10, p. 121018, 2022

    work page 2022

  5. [4]

    Long-term cardiovascular outcomes of COVID-19.,

    Y. Xie, E. Xu, B. Bowe and Z. Al-Aly, "Long-term cardiovascular outcomes of COVID-19.," Nat Med, vol. 28, p. 583–590, 2022

    work page 2022

  6. [5]

    Long-term cardiovascular outcomes of COVID-19.,

    Y. Xie, E. Xu, B. Bowe and Z. Al-Aly, "Long-term cardiovascular outcomes of COVID-19.," Nat Med., vol. 28, no. 3, pp. :583-590, 2022. Information 2025, 16, x ACCEPTED 15 of 17

    work page 2022

  7. [6]

    Núñez-Gil, G

    I. Núñez-Gil, G. Feltes, M. C. Viana-Llamas and e. al., "HOPE-2 Investigators. Post-COVID-19 Symp- toms and Heart Disease: Incidence, Prognostic Factors, Outcomes and Vaccination: Results from a Multi-Center International Prospective Registry (HOPE 2)," J Clin Med., vol. 12, no. 2, p. 706, 2023

    work page 2023

  8. [7]

    Autonomic reactiv- ity to mental stress is associated with cardiovascular mortality

    A. S. Shah, V. Vaccarino, K. Moazzami, Z. Almuwaqqat, M. Garcia, L. Ward, L. Elon, Y-A. Ko, Y. V. Sun, B. D. Pearce, P. Raggi, J. D. Bremner, R. Lampert, A. A. Quyyumi, A. J. Shah. “Autonomic reactiv- ity to mental stress is associated with cardiovascular mortality”, European Heart Journal Open, Vol. 4, Is- sue 6, oeae086, 2024. DOI: https://doi.org/10.10...

    work page doi:10.1093/ehjopen/oeae086 2024

Show all 52 references
  1. [8]

    The Effects of a Single Vagus Nerve’s Neurodynamics on Heart Rate Variability in Chronic Stress: A Randomized Controlled Trial

    A.I. Pérez-Alcalde, F. Galán-del-Río, F.J. Fernández-Rodríguez, M. de la Plaza San Frutos, M. García- Arrabé, M.-J. Giménez, B. Ruiz-Ruiz, “The Effects of a Single Vagus Nerve’s Neurodynamics on Heart Rate Variability in Chronic Stress: A Randomized Controlled Trial”. Sensors,...

    2024 doi

  2. [9]

    An Overview of Heart Rate Variability Metrics and Norms,

    F. Shaffer and J. P. Ginsberg, "An Overview of Heart Rate Variability Metrics and Norms," Frontiers in Public Health, vol. 5, p. 258, 2017

    2017

  3. [10]

    The Importance of Time-domain HRV Analysis in Cardiac Health Predictions,

    S. Chattopadhyay, "The Importance of Time-domain HRV Analysis in Cardiac Health Predictions," Series of Cardiology Research, vol. 4, no. 1, pp. 19-23, 2022

    2022

  4. [11]

    Cardiovascular Risk Prediction Models and Scores in the Era of Personalized Medicine.,

    A. Sofogianni, N. Stalikas, C. Antza and K. Tziomalos, "Cardiovascular Risk Prediction Models and Scores in the Era of Personalized Medicine.," J Pers Med., vol. 12, no. 7, p. 1180, 2022

    2022

  5. [12]

    Observation-Prevention Framework of Cardiac Risk Factors: An Indian Study,

    S. Satapathy and S. Chattopadhyay, "Observation-Prevention Framework of Cardiac Risk Factors: An Indian Study," Journal of Medical Imaging and Health Informatics, vol. 2, no. 2, pp. 102-113(12), 2012

    2012

  6. [13]

    MLMI: A Machine Learning Model for Estimating Risk of Myocardial Infarction,

    S. Chattopadhyay, "MLMI: A Machine Learning Model for Estimating Risk of Myocardial Infarction," Artificial Intelligence Evolution, vol. 5, no. 1, pp. 11-23, 2024

    2024

  7. [14]

    Heart Attack - EDA + Prediction (90% accuracy),

    N. Manchanda, "Heart Attack - EDA + Prediction (90% accuracy)," Kaggle, [Online]. Available: https://www.kaggle.com/code/namanmanchanda/heart-attack-eda-prediction-90-accuracy. [Accessed 19 January 2024]

    2024

  8. [15]

    Coefficient alpha and the internal structure of tests.,

    L. J. Cronbach, "Coefficient alpha and the internal structure of tests.," Psychometrika, vol. 16, no. 3, p. 297–334, 1951

    1951

  9. [16]

    An analysis of variance test for normality (complete samples,

    S. S. Shapiro and M. B. Wilk, "An analysis of variance test for normality (complete samples," Biometrika, vol. 52, no. 3-4, p. 591–611., 1965

    1965

  10. [17]

    K-Means Clustering Algorithm,

    "K-Means Clustering Algorithm," Javapoint, [Online]. Available: https://www.javatpoint.com/k-means- clustering-algorithm-in-machine-learning. [Accessed 21 January 2024]

    2024

  11. [18]

    Gaussian Mixture Models Clustering Algorithm Explained,

    C. Maklin, "Gaussian Mixture Models Clustering Algorithm Explained," Towards Data Science, 16 July

  12. [19]

    One Direction? A Tutorial for Circular Data Analysis Using R With Exam- ples in Cognitive Psychology,

    J. Cremers and I. Klugkist, "One Direction? A Tutorial for Circular Data Analysis Using R With Exam- ples in Cognitive Psychology," Front. Psychol., vol. 9, p. 396689, 2018

    2018

  13. [20]

    DBSCAN Clustering in ML | Density based clustering,

    "DBSCAN Clustering in ML | Density based clustering," geeksforgeeks, [Online]. Available: https://www.geeksforgeeks.org/dbscan-clustering-in-ml-density-based-clustering/. [Accessed 21 Janu- ary 2024]

    2024

  14. [21]

    8 Clustering Algorithms in Machine Learning that All Data Scientists Should Know,

    M. McGregor, "8 Clustering Algorithms in Machine Learning that All Data Scientists Should Know," Freecodecamp, 21 September 2020. [Online]. Available: https://www.freecodecamp.org/news/8-cluster- ing-algorithms-in-machine-learning-that-all-data-scientists-should-know/. [Access...

    2020

  15. [22]

    VIRDOCD: A VIRtual DOCtor to Predict Dengue Fatal- ity,

    A. K. Chattopadhyay and S. Chattopadhyay, "VIRDOCD: A VIRtual DOCtor to Predict Dengue Fatal- ity," Expert Systems, vol. e12796, pp. 1-17, 2021

    2021

  16. [23]

    DBSCAN Python Example: The Optimal Value For Epsilon (EPS),

    C. Maklin, "DBSCAN Python Example: The Optimal Value For Epsilon (EPS)," Towards Data Science, 30 June 2019. [Online]. Available: https://towardsdatascience.com/machine-learning-clustering-dbscan- determine-the-optimal-value-for-epsilon-eps-python-example-3100091cfbc. [Accesse...

    2019

  17. [24]

    B. M. Byrne, Strucrtural equation modeling with AMOS: Basic concepts, applications, and program- ming, New York: Routledge, 2010

    2010

  18. [25]

    Clustering Content Types and User Roles Based on Tweet Text Using K-Medoids Partitioning Based

    R. Benaya, Y. Sibaroni, A. F. Ihsan, “Clustering Content Types and User Roles Based on Tweet Text Using K-Medoids Partitioning Based”, Journal of Computer System and Informatics, Vol. 4. No. 4. Pp. 749-756, 2023

    2023

  19. [27]

    Oxidative stress and cardiovascular disease: new insights,

    P. Pignatelli, D. Menichelli, D. Pastori and F. Violi, "Oxidative stress and cardiovascular disease: new insights," Kardiologia Polska, vol. 76, no. 4, p. 713–722, 2018

    2018

  20. [28]

    The Functional Role of Lipoproteins in Atherosclerosis: Novel Directions for Diagnosis and Targeting Therapy,

    Y. Lu, X. Cui, L. Zhang, X. Wang, Y. Xu, Z. Qin, G. Liu, Q. Wang, K. Tian, K. S. Lim, C. J. Charles, J. Zhang and J. Tang, "The Functional Role of Lipoproteins in Atherosclerosis: Novel Directions for Diagnosis and Targeting Therapy," Aging Dis., vol. 13, no. 2, pp. 491-520, 2022

    2022

  21. [29]

    Uncommon heart attack, found more often in women, needs a second look,

    "Uncommon heart attack, found more often in women, needs a second look," American Heart Associa- tion News, 28 June 2018. [Online]. Available: https://www.heart.org/en/news/2018/07/13/uncommon-heart- attack-found-more-often-in-women-needs-a-second-look. [Accessed 31 January 2024]

    2018

  22. [30]

    Estrogen prevent atherosclerosis by attenuating endothelial cell pyroptosis via activa- tion of estrogen receptor α-mediated autophagy,

    Q. Meng, L. yu, T. ji, Y. Choo, J. Li, Y. Fu, S. Wong, O. Chen, W. Chen, F. Huang, Y. Wang, Q. Zhang, X. Wang and H. Bian, "Estrogen prevent atherosclerosis by attenuating endothelial cell pyroptosis via activa- tion of estrogen receptor α-mediated autophagy," Journal of Advan...

    2021

  23. [31]

    Myocardial Infarction in a Premenopausal Woman on Leuprolide Therapy.,

    I. E. Perez, M. A. Menegus and C. C. Taub, "Myocardial Infarction in a Premenopausal Woman on Leuprolide Therapy.," Case Rep Med. , vol. 2015, no. 390642, 2015

    2015

  24. [32]

    High resting heart rate predicts heart risk in women at midlife,

    "High resting heart rate predicts heart risk in women at midlife," Harvard Health Publishing, 14 May

  25. [33]

    Anxiety disorder in menopausal women and the interven- tion efficacy of mindfulness-based stress reduction,

    S. Huang, Z. Wang, D. Zheng and L. Liu, "Anxiety disorder in menopausal women and the interven- tion efficacy of mindfulness-based stress reduction," Am J Transl Res., vol. 15, no. 3, pp. 2016-2024., 2023

    2016

  26. [34]

    The Role of Estrogen in Anxiety-Like Behavior and Memory of Middle-Aged Female Rats,

    E. Renczés, V. Borbélyová, M. Steinhardt, T. Höpfner, T. Stehle, D. Ostatníková and P. Celec, "The Role of Estrogen in Anxiety-Like Behavior and Memory of Middle-Aged Female Rats," Front Endocrinol (Lau- sanne), vol. 11, p. 570560, 2020

    2020

  27. [35]

    Menopause-associated lipid metabolic disorders and foods beneficial for post- menopausal women,

    S. H. Ko and H. S. Kim, "Menopause-associated lipid metabolic disorders and foods beneficial for post- menopausal women," Nutrients, Vol. 12, no. 202, 2020

    2020

  28. [36]

    Unconventional Monetary Policy and Disaster Risk: Evidence from the Subprime and COVID-19 Crises

    G. S. Cortes, Gao, G. P., Silva, F. B. G., Song, Z. “Unconventional Monetary Policy and Disaster Risk: Evidence from the Subprime and COVID-19 Crises”. Journal of International Money and Finance, Vol. 122, No. 102543, 2022

    2022

  29. [37]

    Corporate Hiring Under COVID -19: Financial Con- straints and the Nature of New Jobs

    M. Campello, G. Kankanhalli, P. Muthukrishnan. “Corporate Hiring Under COVID -19: Financial Con- straints and the Nature of New Jobs”. Journal of Financial and Quantitative Analysis . 2024;59(4):1541- 1585. doi:10.1017/S0022109023000522 Information 2025, 16, x ACCEPTED 17 of 17

    2024 doi

  30. [38]

    High-performance medicine: the convergence of human and artificial intelligence

    E.J. Topol, “High-performance medicine: the convergence of human and artificial intelligence ”. Nat Med Vol. 25, p.44–56, 2019. DOI: https://doi.org/10.1038/s41591-018-0300-7

    2019 doi

  31. [39]

    As artificial intelligence goes multimodal, medical applications multiply

    E.J. Topol, “ As artificial intelligence goes multimodal, medical applications multiply” Science , Vol.381, Issue 6663, 2023. doi: 10.1126/science.adk6139

    2023 doi

  32. [40]

    Adapting to Artificial Intelligence : Radiologists and Pathologists as Information Spe- cialists

    Jha S, Topol E. J. “Adapting to Artificial Intelligence : Radiologists and Pathologists as Information Spe- cialists”. JAMA. 2016;316(22):2353–2354. doi:10.1001/jama.2016.17438

    2016

  33. [41]

    The State of Artificial Intelligence in Precision Oncology: An Interview with Eric Topol

    E. J. Topol, F. Douglas. “The State of Artificial Intelligence in Precision Oncology: An Interview with Eric Topol”, AI in Precision Oncology, vol. 1, no. 1, 2024

    2024

  34. [42]

    A Study on Linkage-based Clustering of Adult Depression Data,

    S. Chattopadhyay, "A Study on Linkage-based Clustering of Adult Depression Data," Journal of Medical Imaging and Health Informatics, vol. 3, pp. 227-236, 2013

    2013

  35. [43]

    Developing Fuzzy Classifiers to Predict the Chance of Occurrence of Adult Psychoses,

    S. Chattopadhyay, D. K. Pratihar and S. C. De Sarkar, "Developing Fuzzy Classifiers to Predict the Chance of Occurrence of Adult Psychoses," Knowledge Based Systems, vol. 21, no. 6, pp. 479-497, 2008

    2008

  36. [44]

    Performance studies of some similarity-based fuzzy clustering algorithms,

    S. Chattopadhyay, D. K. Pratihar and S. C. De Sarkar, "Performance studies of some similarity-based fuzzy clustering algorithms," International Journal of Performability Engineering, vol. 2, no. 2, pp. 191-200, 2007

    2007

  37. [45]

    Comparative Study of Fuzzy k-Nearest Neighbor and Fuzzy C- means Algorithms,

    P. K. Jena and S. Chattopadhyay, "Comparative Study of Fuzzy k-Nearest Neighbor and Fuzzy C- means Algorithms," International Journal of Computer Applications, vol. 57, no. 7, pp. 22-32, 2012

    2012

  38. [46]

    Unsupervised Learning: Clustering using Gaussian Mixture Model (GMM),

    A. Beheshti, "Unsupervised Learning: Clustering using Gaussian Mixture Model (GMM)," Beheshti.me- dium, 11 March 2023. [Online]. Available: https://behesht.medium.com/unsupervised-learning-clustering- using-gaussian-mixture-model-gmm-c788b280932b#:~:text=One%20of%20the%20ad- v...

    2023

  39. [47]

    Extracting prime protein targets as possible drug candidates: machine learning evaluation,

    S. Chattopadhyay, N. P. Do, D. R. Flower and A. K. Chattopadhyay, "Extracting prime protein targets as possible drug candidates: machine learning evaluation," Med Biol Eng Comput, 2023

    2023

  40. [48]

    Towards Predicting Recurrence Risk of Differentiated Thyroid Cancer with a Hy- brid Machine Learning Model

    S. Chattopadhyay, “Towards Predicting Recurrence Risk of Differentiated Thyroid Cancer with a Hy- brid Machine Learning Model” Medinformatics 2024 [Online first]. https://doi.org/10.47852/bonviewME- DIN42024441

    2024 doi

  41. [49]

    Gaussian Mixture Model Implementation for Population Stratification Estimation from Genomics Data,

    A. Budiarto, B. Mahesworo, A. A. Hidayat, I. Nurlaila and B. Pardamean, "Gaussian Mixture Model Implementation for Population Stratification Estimation from Genomics Data," Procedia Computer Science, vol. 179, pp. 202-210, 2021

    2021

  42. [50]

    CHD Risk Minimization through Life- style Control: Machine Learning Gateway

    X. He, R. Matam, S. Bellary, G. Ghosh, A. K. Chattopadhyay, “CHD Risk Minimization through Life- style Control: Machine Learning Gateway”, Scientific Reports 10, 4090, 2020

    2020

  43. [2019]

    Available: https://towardsdatascience.com/gaussian-mixture-models-d13a5e915c8e

    [Online]. Available: https://towardsdatascience.com/gaussian-mixture-models-d13a5e915c8e. [Ac- cessed 21 January 2024]

    2024

  44. [2022]

    Available: https://www.health.harvard.edu/heart-health/in-the-journals-high-resting-heart- rate-predicts-heart-risk-in-women-at-midlife

    [Online]. Available: https://www.health.harvard.edu/heart-health/in-the-journals-high-resting-heart- rate-predicts-heart-risk-in-women-at-midlife. [Accessed 31 January 2024]

    2024

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.