REVIEW 5 major objections 7 minor 1 cited by
CACTUS as a Reliable Tool for Early Classification of Age-related Macular Degeneration
T0 review · 5 major / 7 minor · reviewed 2026-08-15 · deepseek-v4-flash
Pith's one-line read This paper claims that CACTUS, a graph-based explainable classifier, stages age-related macular degeneration from non-imaging data with higher balanced accuracy than six standard machine-learning models across every tested level of…
desk verdict A genuine extension of the authors' CACTUS method with an internally consistent head-to-head win over six ML baselines, but the central comparison may be in-sample because the abstraction and graph parameters are never described as cross-validated per fold. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing object is the per-class knowledge graph: a weighted directed graph whose nodes are abstracted flips (Up and Down for continuous features, allele categories for SNPs) and whose edges are the absolute difference between a conditional flip probability and 0.5, encoding how much more likely two flips are to co-occur than chance. Graph centrality, computed with PageRank or total degree, is multiplied by each flip's class-specific probability to give a significance score; a patient's similarity to a class is the sum of those significances over the flips they show. The new confidence measure, the average absolute difference between the winning class score and the other class scores, is what the paper uses to claim that the PageRank variant is the most trustworthy of the three CACTUS modes.
What would settle it
Train CPR and the same baselines on a dataset built from each patient's first visit, or on a simulated cohort with no dropout, and compare balanced accuracy and top-ranked features; if CPR's advantage over XGBoost shrinks or age loses its top rank, the paper's reliability claim for early screening is not supported. Separately, check whether CPR's confidence values actually order patients by true accuracy within each stage; if they do not, the trust-from-confidence argument fails.
Extended reading notes
Core claim
On the paper's own terms: CACTUS abstracts each continuous feature into Up and Down flips using per-feature ROC cut-offs over all class partitions, builds one weighted directed knowledge graph per AMD stage whose edges are the absolute distance of conditional flip probabilities from 0.5, and scores new patients by combining flip probabilities with graph centrality (Probabilistic, Degree, or PageRank). The central result is that on 29,908 last-visit records from the harmonised EYE-RISK dataset, CACTUS's PageRank variant reaches a balanced accuracy of 0.34 for the five AMD stages, the highest of any model and the best in 3 of 5 fragmentation experiments, while the standard baselines stay between 0.27 and 0.31. Because random guessing is 0.20, the paper interprets these numbers as evidence that the graph representations capture genuine stage-specific structure in non-imaging data, and that the accompanying confidence score lets clinicians decide when to trust a decision.
Load-bearing premise
The results assume that picking each patient's last recorded visit produces a representative and balanced sample of a real early-screening population, so 25 years of dropout does not systematically distort which features predict AMD stage.
Editorial extensions
If this is right
- If CACTUS's PageRank classification holds up, AMD staging can be performed from longitudinal questionnaire and lab data alone, reducing dependence on specialist image grading for screening.
- The confidence metric allows clinicians to defer decisions when confidence is low, concentrating scarce resources on the patients the model is least sure about.
- The feature ranks provide a transparent, testable bridge to medical literature: age topping the ranks and genetic scores appearing near the top are the kinds of outputs clinicians can verify.
- The finding that nine features capture most of the model's performance suggests that a minimal data-collection protocol could be sufficient for early screening.
- Because balanced accuracy stays modest (0.34) even though it beats baselines, the intended use is risk triage to aid a clinician, not autonomous diagnosis.
Reading between the lines
- CPR's edge over XGBoost at higher fragmentation rates suggests the abstraction step may be acting as a noise filter; a direct test would be to compare CACTUS against an XGBoost model trained on discretized features to isolate the contribution of graph centrality.
- The confidence-population curves could be repurposed as a general calibration tool: any classifier can be evaluated by how its accuracy changes when you restrict to high-confidence cases, with CPR serving as a baseline rather than a special case.
- Because the dataset is built from each patient's last visit, the reported 'early classification' performance is measured on the most advanced visit per person; a direct test for early diagnosis would use first visits or time-to-conversion to late AMD as the outcome.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript applies CACTUS, a knowledge-graph-based explainable classifier, to stage age-related macular degeneration (AMD stages 0–4) using non-imaging data from the EYE-RISK/E3 consortium (29,908 patients, 218 features). The method discretizes continuous features via an exhaustive ROC-based abstraction, builds class-specific weighted directed graphs, and assigns labels through three variants: Probabilistic (CPB), Degree (CDG), and PageRank (CPR). The authors compare these variants against six standard ML algorithms (Ridge, RF, LR, SGD, SVM, XGB) at 0%, 20%, 40%, 60%, and 80% artificially induced missing values, reporting that CACTUS variants achieve higher balanced accuracy than all baselines at every fragmentation level (Table 2). They introduce a confidence metric, rank features, and then refine the feature pool and evaluate the top-9 features (Table 3, Figures 4–7). The paper claims that CACTUS, especially CPR, is a reliable explainable tool for early AMD classification.
Significance. If the evaluation is unbiased, the paper demonstrates a transparent, missingness-robust alternative to image-based AMD staging: CACTUS is explainable, provides a confidence metric, and its feature rankings recover established AMD risk factors such as age, ARMS2, CFH, and genetic scores. Comparing against XGBoost is a useful addition. However, absolute balanced accuracy is modest (at most 0.34 versus a 0.20 chance level), the study is purely retrospective with substantial dropout, there is no external validation, and the reported comparisons carry a material risk of in-sample evaluation. The practical utility for early screening is therefore not established by this manuscript alone.
major comments (5)
- [Methods: 'Abstractions', 'The Knowledge Graphs', and 'Classification'; Figure 1] The paper does not state whether CACTUS's fitted components—the exhaustive ROC-based cut-off search per continuous feature, the class-partition selection, and the conditional-probability graph weights—are recomputed inside each training fold. The Classification section and Figure 1 say the model is built and 'the classification is then applied to all the available individuals to assess the performance,' while the ML baselines are described with an 80/20 split and 10-fold cross-validation. If the flips and graph parameters are estimated on the full data before scoring, the test labels have influenced the feature discretization and graph structure, so the Table 2 advantage (e.g., CPR 0.34 vs XGB 0.31 at 0% missingness) could be leakage-driven. The authors must either provide a nested evaluation in which all fitted parameters are estimated on training folds only and applied to held-out folds, or otherwise demonstrate that the comparison is not in-sample.
- [Results: 'CACTUS ranks'; Table 3] The refinement step selects the nine highest-ranked features from ranks computed on the full dataset and then reports balanced accuracy on the same data. Because feature selection and evaluation share the same samples, the Table 3 results (CDG 31%, CPB 29%, CPR 29% on the nine highest ranks) are at risk of selection bias and cannot be interpreted as unbiased performance estimates. The authors should perform feature selection inside an inner cross-validation loop or on a separate validation cohort, and then evaluate on an untouched test portion.
- [Methods: 'The defined dataset' paragraph and Table 1] The dataset is constructed from the last available visit per patient ('To have both a populous and balanced dataset, we considered the last visit available for each person'), and Table 1 shows heavy dropout over the study (from 28,910 participants at visit 1 to 1,538 at visit 5, with 17,526 dropped by visit 2). If dropout is correlated with disease severity, health status, or age, the resulting cohort may not represent an early-screening population, and the reported accuracy and the top ranking of age as a feature may be inflated. The manuscript should discuss this selection mechanism and its possible directional bias, or provide sensitivity analyses stratified by visit completeness.
- [Results: Table 2 and Discussion] The central performance comparison is reported only as mean ± standard deviation, with no statistical significance tests, confidence intervals, or paired comparisons across cross-validation folds. Given the small absolute differences (e.g., 0.34±0.01 vs 0.31±0.01 at 0% missingness) and overlapping standard deviations, the claim that CACTUS 'outperforms' the baselines is not statistically supported by the evidence as presented. Please add appropriate significance testing, such as paired tests on the cross-validation folds or bootstrap confidence intervals for the differences.
- [Results: 'CACTUS ranks' versus Table 3] The text states 'This filtration decreased the balanced accuracy by 0% to 2% for CDG, CPR, and CPB, as shown in Table 3,' but Table 3 reports CDG dropping from 33% (whole dataset) to 26% (refined dataset), a seven-percentage-point decrease. This internal inconsistency between the stated claim and the displayed results must be corrected and, if the refined-dataset CDG value is accurate, the discussion of the refinement step should be revised accordingly.
minor comments (7)
- [Figure 2 legend] The figure legend labels the model 'Linear Regression (LR),' while the Methods section refers to 'Logistic Regression (LR);' the labels should be made consistent.
- [Methods: 'Comparison against standard Machine Learning algorithms'] A stray character 'Z' appears after 'Logistic Regression (LR)' in the text; please remove it.
- [Table 2 caption] The caption says the balanced accuracy is reported 'in percentage,' but the table entries are decimals (e.g., 0.33, not 33); either the caption or the entries should be adjusted for consistency.
- [Results: 'CACTUS confidence'] The phrase 'The more inflated the cumulative lines for confidence and populations, the better the corresponding model' is difficult to parse and should be rewritten for clarity.
- [Results: 'CACTUS confidence' and Figure 3] The text mentions a 'confidence threshold of 15%' that 'enables us to confidently apply CPR on 50% of the population,' but Figure 3 annotates vertical lines as the confidence levels covering 90%, 80%, 70%, 60%, and 50% of the population; please clarify which line corresponds to the 15% threshold or adjust the text.
- [Methods: 'Comparison against standard Machine Learning algorithms'] The statement that CACTUS 'does not make any assumptions (weak model) on how the real process works' is vague, and the citations to White and Hellström do not directly justify the overfitting argument; please provide a more precise complexity comparison (e.g., number of effective parameters versus training sample size).
- [General] No data or code availability statement is provided; a clear statement on code and trained model availability would be important for reproducibility of the main comparison.
Circularity Check
No significant circularity: the CACTUS equations are self-contained and the listed self-citation is not load-bearing; the in-sample evaluation concern is a protocol-transparency issue, not a demonstrated circularity.
full rationale
The paper's derivation chain is not circular in the definitional sense. Equations (1)-(4) define conditional-probability graph weights, centrality, rank, and confidence from the data; none of them assumes the conclusion that CACTUS classifies AMD stages well. The abstraction thresholds are chosen by ROC search over the classes, but this is a supervised fitting procedure, not a circular prediction; if the evaluation were held out, the fitted values would not force the reported balanced accuracy. The only possible circularity would be evaluating CACTUS on the same data used to build its graphs. The text does contain a sentence in Figure 1 saying the classification is applied to all available individuals, but Table 2 explicitly states that all models were tested with a 10-fold cross-validation and an 80/20 split, so the paper as a whole does not unambiguously demonstrate a resubstitution-only evaluation. An underspecified CACTUS-specific CV protocol is a reproducibility and correctness risk, not a shown circular step. The citation of the authors' prior CACTUS paper [23] is a normal self-citation: the current manuscript re-describes the algorithm in sufficient detail (abstractions, knowledge graphs, centrality, rank, confidence), so the central claim does not rest on the cited paper's authority. No uniqueness theorem, ansatz-by-citation, or renaming of a known result is present.
Assumptions & free parameters
free parameters (3)
- Per-feature ROC cut-off thresholds (Up/Down flips) =
One optimal threshold per continuous feature, not listed in paper
- Class partition choice per feature during abstraction =
Exhaustive search over all class partitions, not tabulated
- 50% missing-value feature exclusion threshold =
50%
assumptions (4)
- domain assumption AMD stage labels derived from retinal image grading are treated as ground truth.
- domain assumption The last available visit per patient is a valid cross-sectional representation for screening classification.
- standard math Standard definitions of conditional probability, graph centrality, PageRank, and balanced accuracy apply as implemented.
- domain assumption Randomly-induced missingness behaves like real-world missingness.
Cite this review
Pith. "Pith review of CACTUS as a Reliable Tool for Early Classification of Age-related Macular Degeneration." pith.science (2026). https://pith.science/paper/4S3LF5FU
@misc{pith2026250614843,
author = {Pith},
title = {Pith review of: CACTUS as a Reliable Tool for Early Classification of Age-related Macular Degeneration},
year = {2026},
howpublished = {\url{https://pith.science/paper/4S3LF5FU}},
note = {Machine review of arXiv:2506.14843}
}
read the original abstract
Machine Learning (ML) is used to tackle various tasks, such as disease classification and prediction. The effectiveness of ML models relies heavily on having large amounts of complete data. However, healthcare data is often limited or incomplete, which can hinder model performance. Additionally, issues like the trustworthiness of solutions vary with the datasets used. The lack of transparency in some ML models further complicates their understanding and use. In healthcare, particularly in the case of Age-related Macular Degeneration (AMD), which affects millions of older adults, early diagnosis is crucial due to the absence of effective treatments for reversing progression. Diagnosing AMD involves assessing retinal images along with patients' symptom reports. There is a need for classification approaches that consider genetic, dietary, clinical, and demographic factors. Recently, we introduced the -Comprehensive Abstraction and Classification Tool for Uncovering Structures-(CACTUS), aimed at improving AMD stage classification. CACTUS offers explainability and flexibility, outperforming standard ML models. It enhances decision-making by identifying key factors and providing confidence in its results. The important features identified by CACTUS allow us to compare with existing medical knowledge. By eliminating less relevant or biased data, we created a clinical scenario for clinicians to offer feedback and address biases.
Figures
Figures from the paper (4 more)
Forward citations
Cited by 1 Pith paper
-
CLARK: Closed-loop Learning for Adaptive Reasoning over Knowledge Graphs
CLARK improves medical classification by iteratively enriching a knowledge graph with mined logical rules calibrated through LP^MLN probabilistic reasoning.
Reference graph
Works this paper leans on
-
[1]
Adadi, A survey on data-efficient algorithms in big data era, Vol
A. Adadi, A survey on data-efficient algorithms in big data era, Vol. 8, Springer International Pub- lishing, 2021. doi:10.1186/s40537-021-00419-9 . URL https://doi.org/10.1186/s40537-021-00419-9
-
[2]
J. Lee, C. Liu, J. Kim, Z. Chen, Y . Sun, J. R. Rogers, W. K. Chung, C. Weng, Deep learning for rare disease: A scoping review, Journal of Biomedical Informatics 135 (2022) 104227.doi:https: //doi.org/10.1016/j.jbi.2022.104227. URL https://www.sciencedirect.com/science/article/pii/S1532046422002325
-
[3]
F . Wang, A. Preininger, AI in Health: State of the Art, Challenges, and Future Directions, Y earbook of medical informatics 28 (1) (2019) 16–26. doi:10.1055/s-0039-1677908
-
[4]
G. S. Handelman, H. K. Kok, R. V. Chandra, A. H. Razavi, S. Huang, M. Brooks, M. J. Lee, H. As- adi, Peering Into the Black Box of Artificial Intelligence: Evaluation Metrics of Machine Learning Methods, American Journal of Roentgenology 212 (1) (2019) 38–43.doi:10.2214/AJR.18.20224. URL http://services.igi-global.com/resolvedoi/resolve.aspx?doi=10.4018/ ...
-
[5]
A. J. London, Artificial Intelligence and Black-Box Medical Decisions: Accuracy versus Explainab- ility, Hastings Center Report 49 (1) (2019) 15–21. doi:10.1002/hast.973
doi:10.1002/hast.973 2019
-
[6]
A. Das, P . Rad, Opportunities and Challenges in Explainable Artificial Intelligence (XAI): A Survey, Diabetes Care 41 (2020) 1–24. arXiv:2006.11371. URL http://arxiv.org/abs/2006.11371
arXiv 2020
-
[7]
P . P . Angelov, E. A. Soares, R. Jiang, N. I. Arnold, P . M. Atkinson, Explainable artificial intelligence: an analytical review, Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery 11 (5) (2021) 1–13. doi:10.1002/widm.1424
-
[8]
R. Miotto, F . Wang, S. Wang, X. Jiang, J. T. Dudley, Deep learning for healthcare: review, op- portunities and challenges, Briefings in Bioinformatics 19 (6) (2017) 1236–1246. arXiv:https: //academic.oup.com/bib/article-pdf/19/6/1236/27119191/bbx044.pdf, doi:10.1093/bib/ bbx044. URL https://doi.org/10.1093/bib/bbx044
doi:10.1093/bib/ 2017
Show all 50 references
-
[9]
A. S. Adamson, A. Smith, Machine Learning and Health Care Disparities in Dermatology, JAMA Dermatology 154 (11) (2018) 1247. doi:10.1001/jamadermatol.2018.2348. URL http://archderm.jamanetwork.com/article.aspx?doi=10.1001/jamadermatol.2018. 2348
2018
-
[10]
M. Du, F . Y ang, N. Zou, X. Hu, Fairness in deep learning: A computational perspective, IEEE Intelligent Systems 36 (4) (2021) 25–34. doi:10.1109/MIS.2020.3000681. 21
2021
-
[11]
MacCarthy, An Examination of the Algorithmic Accountability Act of 2019, SSRN Electronic Journal (2020) 1–10doi:10.2139/ssrn.3615731
M. MacCarthy, An Examination of the Algorithmic Accountability Act of 2019, SSRN Electronic Journal (2020) 1–10doi:10.2139/ssrn.3615731
2020 doi
-
[12]
Right to Explanation
Goodman Bryce, Flaxman Seth, European Union Regulations on Algorithmic Decision Making and a “Right to Explanation”, AI magazine 38 (3) (2017) 50–57
2017
-
[13]
Guidotti, A
R. Guidotti, A. Monreale, D. Pedreschi, F . Giannotti, Principles of Explainable Artificial Intelligence, Explainable AI Within the Digital Transformation and Cyber Physical Systems (2021) 9–31 doi: 10.1007/978-3-030-76409-8_2
2021 doi
-
[14]
X.-H. Li, C. C. Cao, Y . Shi, W. Bai, H. Gao, L. Qiu, C. Wang, Y . Gao, S. Zhang, X. Xue, L. Chen, A Survey of Data-driven and Knowledge-aware eXplainable AI, IEEE Transactions on Knowledge and Data Engineering 34 (1) (2022) 29–49. doi:10.1109/TKDE.2020.2983930. URL https://ie...
2022
-
[15]
Tiddi, S
I. Tiddi, S. Schlobach, Knowledge graphs as tools for explainable machine learning: A survey, Ar- tificial Intelligence 302 (2022) 103627. doi:https://doi.org/10.1016/j.artint.2021.103627. URL https://www.sciencedirect.com/science/article/pii/S0004370221001788
2022
- [16]
-
[17]
D. N. Nicholson, C. S. Greene, Constructing knowledge graphs and their biomedical applications, Computational and Structural Biotechnology Journal 18 (2020) 1414–1428. doi:https://doi. org/10.1016/j.csbj.2020.05.017. URL https://www.sciencedirect.com/science/article/pii/S20010...
2020 doi
-
[18]
W. L. Wong, X. Su, X. Li, C. M. G. Cheung, R. Klein, C. Y . Cheng, T. Y . Wong, Global pre- valence of age-related macular degeneration and disease burden projection for 2020 and 2040: A systematic review and meta-analysis, The Lancet Global Health 2 (2) (2014) e106–e116. doi:...
2014 doi
-
[19]
´Spiewak, Ł
D. ´Spiewak, Ł. Drzyzga, M. Dorecka, D. Wygle ¸dowska-Promie´nska, Summary of the Therapeutic Options for Patients with Dry and Neovascular AMD, Journal of Clinical Medicine 13 (14) (2024). doi:10.3390/jcm13144227
2024 doi
-
[20]
C. N. Thomas, D. A. Sim, W. H. Lee, N. Alfahad, A. D. Dick, A. K. Denniston, L. J. Hill, Emer- ging therapies and their delivery for treating age-related macular degeneration, British Journal of Pharmacology 179 (9) (2022) 1908–1937. doi:10.1111/bph.15459. 22
2022 doi
-
[21]
E. Y . Chew, Age-related Macular Degeneration: Nutrition, Genes and Deep Learning—The LXXVI Edward Jackson Memorial Lecture, American Journal of Ophthalmology 217 (2020) 335–347.doi: 10.1016/j.ajo.2020.05.042. URL https://doi.org/10.1016/j.ajo.2020.05.042
2020 doi
-
[22]
X. Liu, L. Faes, A. U. Kale, S. K. Wagner, D. J. Fu, A. Bruynseels, T. Mahendiran, G. Moraes, M. Shamdas, C. Kern, J. R. Ledsam, M. K. Schmid, K. Balaskas, E. J. Topol, L. M. Bachmann, P . A. Keane, A. K. Denniston, A comparison of deep learning performance against health-care...
2019 doi
-
[23]
Gherardini, V
L. Gherardini, V. R. Varma, K. Capała, R. Woods, J. Sousa, CACTUS: a Comprehensive Abstrac- tion and Classification Tool for Uncovering Structures, ACM Transactions on Intelligent Systems and Technology (feb 2024). arXiv:2308.12031, doi:10.1145/3649459. URL https://dl.acm.org/...
2024 arXiv
-
[24]
Garnelo, M
M. Garnelo, M. Shanahan, Reconciling deep learning with symbolic artificial intelligence: repres- enting objects and relations, Current Opinion in Behavioral Sciences 29 (2019) 17–23, artificial Intelligence. doi:https://doi.org/10.1016/j.cobeha.2018.12.010. URL https://www.sc...
2019 doi
-
[25]
Salimiaghdam, M
N. Salimiaghdam, M. Riazi-Esfahani, P . S. Fukuhara, K. Schneider, M. C. Kenney, Age-related Macular Degeneration (AMD): A Review on its Epidemiology and Risk Factors, The Open Ophthal- mology Journal 13 (1) (2020) 90–99. doi:10.2174/1874364101913010090
2020 doi
-
[26]
T. D. Keenan, S. Vitale, E. Agr ´on, A. Domalpally, A. N. Antoszyk, M. J. Elman, T. E. Clemons, E. Y . Chew, Visual Acuity Outcomes after Anti–Vascular Endothelial Growth Factor Treatment for Neovascular Age-Related Macular Degeneration, Ophthalmology Retina 4 (1) (2020) 3–12....
2020 doi
-
[27]
C.-J. Chiu, R. Klein, R. C. Milton, G. Gensler, A. Taylor, Does eating particular diets alter the risk of age-related macular degeneration in users of the age-related eye disease study supple- ments?, British Journal of Ophthalmology 93 (9) (2009) 1241–1246. arXiv:https://bjo....
2009
-
[28]
Q. Y an, D. E. Weeks, H. Xin, A. Swaroop, E. Y . Chew, H. Huang, Y . Ding, W. Chen, Deep-learning- based prediction of late age-related macular degeneration progression, Nature Machine Intelli- gence 2 (2) (2020) 141–150. doi:10.1038/s42256-020-0154-9 . URL http://dx.doi.org/1...
2020 doi
-
[29]
P . M. Burlina, N. Joshi, M. Pekala, K. D. Pacheco, D. E. Freund, N. M. Bressler, Automated grading of age-related macular degeneration from color fundus images using deep convolutional neural net- works, JAMA Ophthalmology 135 (11) (2017) 1170–1176. doi:10.1001/jamaophthalmol...
2017 doi
-
[30]
P . M. Burlina, N. Joshi, K. D. Pacheco, D. E. Freund, J. Kong, N. M. Bressler, Use of Deep Learning for Detailed Severity Characterization and Estimation of 5-Y ear Risk among Patients with Age- Related Macular Degeneration, JAMA Ophthalmology 136 (12) (2018) 1359–1366. doi:1...
2018
-
[31]
Schmidt-Erfurth, A
U. Schmidt-Erfurth, A. Sadeghipour, B. S. Gerendas, S. M. Waldstein, H. Bogunovi ´c, Artificial intelligence in retina, Progress in Retinal and Eye Research 67 (May) (2018) 1–29. doi:10.1016/ j.preteyeres.2018.07.004. URL https://doi.org/10.1016/j.preteyeres.2018.07.004
2018 doi
-
[32]
Michl, M
M. Michl, M. Fabianska, P . Seeb ¨ock, A. Sadeghipour, B. H. Najeeb, H. Bogunovic, U. M. Schmidt-Erfurth, B. S. Gerendas, Automated quantification of macular fluid in retinal diseases and their response to anti-VEGF therapy, British Journal of Ophthalmology (2020) 1–8 doi: 10....
2020 doi
-
[33]
Perepelkina, A
T. Perepelkina, A. B. Fulton, Artificial Intelligence (AI) Applications for Age-Related Macular Degen- eration (AMD) and Other Retinal Dystrophies, Seminars in Ophthalmology 36 (4) (2021) 304–309. doi:10.1080/08820538.2021.1896756. URL https://doi.org/10.1080/08820538.2021.1896756
2021
-
[34]
D. S. W. Ting, C. Y . L. Cheung, G. Lim, G. S. W. Tan, N. D. Quang, A. Gan, H. Hamzah, R. Garcia- Franco, I. Y . S. Y eo, S. Y . Lee, E. Y . M. Wong, C. Sabanayagam, M. Baskaran, F . Ibrahim, N. C. Tan, E. A. Finkelstein, E. L. Lamoureux, I. Y . Wong, N. M. Bressler, S. Sivapr...
2017
-
[35]
Bogunovi ´c, S
H. Bogunovi ´c, S. M. Waldstein, T. Schlegl, G. Langs, A. Sadeghipour, X. Liu, B. S. Gerendas, A. Osborne, U. Schmidt-Erfurth, Prediction of Anti-VEGF Treatment Requirements in Neovascular AMD Using a Machine Learning Approach, Investigative Ophthalmology and Visual Science 58...
2017 doi
-
[36]
European Eye Epidemiology
C. Delcourt, J. F . Korobelnik, G. H. Buitendijk, P . J. Foster, C. J. Hammond, S. Piermarocchi, T. Peto, N. Jansonius, A. Mirshahi, R. E. Hogg, L. Bretillon, F . Topouzis, G. Deak, J. Grauslund, R. Broe, E. H. Souied, C. Creuzot-Garcher, J. Sahel, V. Daien, T. Lehtim ¨aki, H....
2016
-
[37]
T. P . Peixoto, The graph-tool python library, figshare (2014).doi:10.6084/m9.figshare.1164194
2014 doi
-
[38]
D. F . Gleich, Pagerank beyond the web, SIAM Review 57 (3) (2015) 321–363. arXiv:https: //doi.org/10.1137/140976649, doi:10.1137/140976649. URL https://doi.org/10.1137/140976649
2015 doi
-
[39]
Nouretdinov, S
I. Nouretdinov, S. G. Costafreda, A. Gammerman, A. Chervonenkis, V. Vovk, V. Vapnik, C. H. Fu, Machine learning classification with confidence: Application of transductive conformal predictors to mri-based diagnostic and prognostic markers in depression, NeuroImage 56 (2) (201...
2011 doi
-
[40]
T. A. Johansen, On tikhonov regularization, bias and variance in nonlinear system identification, Automatica 33 (3) (1997) 441–446. doi:https://doi.org/10.1016/S0005-1098(96)00168-9
1997 doi
-
[41]
M. P . LaValley, Logistic regression, Circulation 117 (18) (2008) 2395–2399. doi:10.1161/ CIRCULATIONAHA.106.682658
2008
-
[42]
Cortes, V
C. Cortes, V. Vapnik, Support-vector networks, Machine Learning 20 (3) (1995) 273–297. doi: 10.1007/BF00994018
1995 doi
-
[43]
ichi Amari, Backpropagation and stochastic gradient descent method, Neurocomputing 5 (4) (1993) 185–196
S. ichi Amari, Backpropagation and stochastic gradient descent method, Neurocomputing 5 (4) (1993) 185–196. doi:https://doi.org/10.1016/0925-2312(93)90006-O
1993 doi
-
[44]
Breiman, Random Forests, Machine Learning 45 (2001) 5–32.doi:https://doi.org/10.1023/ A:1010933404324
L. Breiman, Random Forests, Machine Learning 45 (2001) 5–32.doi:https://doi.org/10.1023/ A:1010933404324
2001
-
[45]
T. Chen, C. Guestrin, Xgboost: A scalable tree boosting system, in: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD ’16, Association for Computing Machinery, New Y ork, NY , USA, 2016, p. 785–794.doi:10.1145/2939672.29397...
2016
-
[46]
Van Rossum, F
G. Van Rossum, F . L. Drake, Python 3 Reference Manual, CreateSpace, Scotts Valley, CA, 2009
2009
-
[47]
Pedregosa, G
F . Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, P . Pretten- hofer, R. Weiss, V. Dubourg, J. Vanderplas, A. Passos, D. Cournapeau, M. Brucher, M. Perrot, E. Duchesnay, Scikit-learn: Machine learning in Python, Journal of Machine Learning...
2011
-
[48]
Ying, An Overview of Overfitting and its Solutions, Journal of Physics: Conference Series 1168 (2) (2019)
X. Ying, An Overview of Overfitting and its Solutions, Journal of Physics: Conference Series 1168 (2) (2019). doi:10.1088/1742-6596/1168/2/022022
2019 doi
-
[49]
H. White, Parametric statistical estimation with artificial neural networks: A condensed discussion”, booktitle=”from statistics to neural networks, Springer Berlin Heidelberg, Berlin, Heidelberg, 1994, pp. 127–146
1994
-
[50]
Hellstr ¨om, A Random Walk through the Stock Market, Doctoral dissertation (1998)
T. Hellstr ¨om, A Random Walk through the Stock Market, Doctoral dissertation (1998). URL http://e-m-h.org/Hell98.pdf 26 Supplementary Figure 5: Feature ranks computed by CACTUS through the Degree centrality on the knowledge graphs. They are sorted from the highest (upper left...
1998
Reviewed August 15, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.