PolyModel for Hedge Funds' Portfolio Construction Using Machine Learning

Dan Wang; Raphael Douady; Siqiao Zhao

REVIEW 5 major objections 4 minor 1 cited by

PolyModel for Hedge Funds' Portfolio Construction Using Machine Learning

T0 review · 5 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Machine-learning fund selection with all PolyModel filters lifts cumulative returns

desk verdict A practical hedge-fund selection study with a plausible but weakly supported central claim; the missing-data imputation is the biggest problem and it is not addressed. read the letter →

arxiv 2412.11019 v1 pith:QRUHYAP6 submitted 2024-12-15 q-fin.ST q-fin.PM

Siqiao Zhao , Dan Wang , Raphael Douady This is my paper · ORCID

classification q-fin.STq-fin.PM

keywords hedgefundsmachinelearningXGBoostPolyModeltheoryfeatureselectionportfolioconstructionfundsizeLong-TermStability

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

This paper claims that a hedge fund portfolio strategy gains from combining machine learning with PolyModel feature filters: using XGBoost to predict which funds will have positive returns next month, then keeping only funds that pass all three risk-adjusted filters (LTS, MRaR, Sharpe), produces higher average cumulative returns than using no machine learning or fewer filters. The paper also claims that equal-weighting the selected funds beats weighting by assets under management, so fund size alone is not a reliable signal of performance. A sympathetic reader would care because the recipe is concrete and testable: it suggests a data-driven replacement for size-based fund selection.

What carries the argument

The load-bearing mechanism is the PolyModel feature-generation stage feeding an XGBoost classifier. PolyModel regresses each fund's returns on a pool of risk factors using degree-4 Hermite polynomials, uses target shuffling to compute P-value scores of factor importance, and builds tail-aware features including StressVaR and Long-Term Stability. Those features, plus monthly return and AUM, train XGBoost to output the probability of a positive next-month return; funds are then kept only if their LTS, MRaR, and Sharpe values clear the paper's thresholds, and cash is split evenly across survivors.

What would settle it

Recompute the backtests with missing monthly returns imputed by a neutral method (e.g., median return or cross-sectional mean) or by excluding missing months, and compare the ML versus no-ML cumulative returns; if the gap shrinks or reverses, the central claim fails.

Watch

Extended reading notes

Core claim

The paper's central discovery, stated on its own terms, is that the full pipeline—XGBoost direction predictions plus all three PolyModel filters, with equal weights—is the best configuration in the backtest, reaching a cumulative return of about 41.8. It reports that machine learning raises average cumulative return (about 24.1 vs 19.6 for no-ML) at the cost of higher annual volatility, that using all three filters dominates using fewer, and that Long-Term Stability alone gives a strong Sharpe ratio (about 1.65) while controlling drawdown. Table 5 is read as evidence that AUM-weighted allocation (15.6 average cumulative return) underperforms equal allocation (28.1), challenging the idea that larger funds are more reliable.

Load-bearing premise

The paper's comparisons all depend on the assumption that filling missing monthly returns with -30 and missing Sharpe, LTS, and MRaR values with -3, -1, and -3 does not systematically distort which funds get selected.

Editorial extensions

If this is right

Portfolios built with machine-learning fund selection realize larger cumulative returns than no-ML portfolios, but with higher annual volatility.
Using all three PolyModel filters (LTS, Sharpe, MRaR) yields higher cumulative returns than using any single filter or none; the best configuration reaches a cumulative return of 41.8.
Long-Term Stability as a standalone filter produces a high average Sharpe ratio with a small max drawdown, making it a useful volatility-control feature.
Equal-weight allocation across selected funds outperforms AUM-weighting on average cumulative return, so fund size should not be the basis for hedge fund portfolio weighting.

Reading between the lines

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

A likely artifact risk is the fixed missing-data imputation: if funds with missing monthly returns are systematically weaker, the -30 fill could manufacture the appearance that filters and ML select better funds.
Since the backtest assumes zero transaction costs and monthly rebalancing across many funds, the reported cumulative-return gap would likely narrow under realistic costs; testing with a small cost model would quantify this.
The same pipeline—ML direction prediction plus multiple risk-adjusted filters and equal weighting—could be applied to mutual fund or managed-account datasets to see whether the conclusion generalizes.

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.

Desk Editor's Note

read the letter

Quick take: this is an incremental empirical paper in a line of work the authors have already published. The new ingredient is XGBoost as the selector, plus a comparison of equal-weight vs AUM-weight allocation. The paper is not a breakthrough, but it is not empty either. The central claims are plausible: ML-based selection improves cumulative returns, comprehensive filters help, and AUM-weighting does not beat equal-weight. The problem is that the evidence as reported does not support these claims with any statistical rigor.

What is actually new: the use of XGBoost on PolyModel features for hedge fund selection, with a specific set of filters (LTS, MRaR, Sharpe) and a comparison across 10,545 funds. The description of the PolyModel construction and the feature definitions is clear enough to replicate the procedure, though no code or data are provided. The equal-weight vs AUM-weight finding, if it holds, is a useful caution against defaulting to size-based allocation.

The soft spots are serious and load-bearing. Section 3.2 fills missing returns with -30, Sharpe with -3, LTS with -1, MRaR with -3, and the dataset has 'a significant number of missing values.' Hedge fund reporting is voluntary; missingness is likely correlated with fund age, liquidation, or reporting quality. With these fills, any non-reporting fund looks catastrophically bad, and both the XGBoost model and the filters will learn to exclude those funds. The observed patterns—ML improves returns, more filters help, equal-weight beats AUM-weight—could be entirely driven by this artifact. There is no sensitivity analysis, no alternative imputation, no missingness indicator, and no limitation statement. That is a serious omission.

Second, Tables 3-5 report averages across experiments with no error bars, no significance tests, and no indication of how many experiments each average covers. The best performer in Table 6 is selected ex post, so its numbers are cherry-picked by construction. Filter thresholds are chosen from the same historical data used for evaluation, which invites lookahead bias. And LTS is defined as LTA - κ·SVaR, so reporting that LTS filtering 'controls volatility' is partly a restatement of the construction rather than an empirical discovery.

Who this is for: practitioners in quantitative asset management who might want to test the missing-data fix and see whether the equal-weight result survives. An academic reader will be frustrated by the lack of inference. It deserves a serious referee—I would send it to review, not desk-reject—but the referee should ask for a major revision that addresses the imputation, adds significance testing, and moves the threshold selection out-of-sample.

Referee Report

5 major / 4 minor

Summary. The paper proposes a machine-learning pipeline for hedge fund portfolio construction in which PolyModel-derived features (LTS, MRaR, Sharpe ratio, monthly return, and AUM) are fed into XGBoost to predict the direction of next-month returns, and funds passing threshold filters on LTS, MRaR, Sharpe, and predicted probability are selected for equal- or AUM-weighted portfolios. Using monthly data on 10,545 hedge funds from April 1994 to May 2023, the paper claims that machine learning increases cumulative returns, that comprehensive PolyModel feature filters outperform partial or no filters, and that equal-weighting beats AUM-weighting. The manuscript also highlights a 'best performer' configuration that combines all three filters, machine learning, and equal weighting, with a cumulative return of 41.81.

Significance. If the empirical claims were properly supported, the practical implications would be meaningful for fund-of-funds and asset allocators: a reproducible pipeline combining nonlinear factor features with tree-based direction prediction could improve fund selection and challenge size-based allocation heuristics. The paper is transparent about some construction choices, such as the explicit definition of LTS in Section 2.2.6 and the imputation constants in Section 3.2, and it addresses a real problem of sparse hedge fund reporting. However, the manuscript provides no code or data, no error bars or significance tests for the headline tables, no sensitivity analysis for the missing-data imputation, and no validation that the filter thresholds are chosen without look-ahead. As it stands, the significance is prospective rather than demonstrated.

major comments (5)

[§3.2] The imputation of missing monthly returns with -30, Sharpe with -3, LTS with -1, and MRaR with -3 is asserted rather than justified, even though the paper itself states that the dataset has a significant number of missing values. Because hedge fund reporting is voluntary, missingness is plausibly correlated with fund age, size, liquidation, or reporting quality, and the punitive fill values will mechanically push non-reporting funds below every filter threshold and into the XGBoost negative class. No sensitivity analysis, alternative imputation, or missingness-indicator test is reported, so all three headline comparisons (ML vs. no ML, full vs. partial filters, equal vs. AUM weighting) could be artifacts of this imputation rather than effects of the methods.
[§3.4.2–§3.4.4, Tables 3–5] The central claims are supported only by point estimates of mean performance with no standard errors, confidence intervals, or significance tests. For example, Table 3 reports cumulative returns of 19.59 versus 24.05 and Sharpe ratios of 1.2006 versus 1.1819, but without dispersion measures it is impossible to tell whether either difference is meaningful; the same problem affects Table 5's cumulative-return gap of 28.09 versus 15.56. The paper's language that the data 'decisively address' or 'unequivocally show' the conclusions is therefore not supported by the reported evidence.
[§2.3, §3.3] The filter thresholds for LTS, MRaR, Sharpe, and the predicted probability p_i are said to be determined based on empirical analysis and historical market performance, but the text does not state whether these thresholds were selected on data prior to the evaluation period or on the same full sample. If the latter, the filtering and the XGBoost predictions are effectively evaluated in-sample, and the reported outperformance of the comprehensive-filter strategies would be an artifact of threshold selection rather than a predictive edge.
[§2.2.6, Table 4] The claimed volatility-control benefit of LTS is partly mechanical: Section 2.2.6 defines LTS as LTA - κ·SVaR with κ = 5%, so funds with high LTS are constructed to have low SVaR. Table 4's LTS-only row does show lower max drawdown (0.089) and higher Sharpe (1.650) than no filtering, but the comprehensive-filter row (LTS, Sharpe, MRaR) raises max drawdown to 0.338 while lowering Sharpe to 1.184, which undercuts the conclusion that using all available features is consistently better. The paper needs a risk-adjusted comparison that accounts for the construction-induced correlation between LTS and volatility.
[§3.4.5, Table 6] The best-performer showcase is selected ex post from a grid of experiments (ML × filters × weighting), yet it is presented as validation of the approach. Without a multiple-testing correction, a held-out period, or at least the full distribution of outcomes over the grid, the best-cell result of 41.81 cumulative return is expected to be inflated by selection and cannot be used as evidence for the method's effectiveness.

minor comments (4)

[§2.1, §2.1.2, §2.2.3] The text contains several typos and cross-reference errors: 'PloyModel' should be 'PolyModel', 'chanllenges' should be 'challenges', 'detials' should be 'details', 'non-casual' should be 'non-causal', and the reference to 'Section 2.5.2' in Section 2.2.3 should be to Section 2.1.3.
[Tables 1–2] Table 1 is captioned 'List of the Risk Factors for Hedge Funds Portfolio Construction' but shows hedge funds, while Table 2 is captioned 'List of the Risk Factors for Network' and shows benchmark indices; the captions are confusing and should be corrected.
[§3.1, §3.3] The paper states that the dataset covers 10,545 funds from April 1994 to May 2023, but the experimental section does not report how many funds have sufficient history for the two-year training window, how the rolling window is initialized, or what the actual out-of-sample evaluation period is.
[§3.4.1] Several metrics listed in Section 3.4.1, such as 'Average Monthly Increase' and 'Correlations with HFRIFOF/HFRIFWI', are defined but not used in the interpretation of Tables 3–5; the authors should either use them in the analysis or remove them from the list.

Circularity Check

2 steps flagged · score 6.0 of 10

LTS's risk-control 'finding' is partly built into its definition, and filter thresholds are tuned on the same historical data used to claim that comprehensive filters outperform.

self definitional [Section 2.2.6 and Section 3.4.3]
"For fund Yi, LT Si := LT Ai − κ · SV aRi where κ is a hyper-parameter whose value is set to 5%. ... A particularly notable finding from this study is the role of LTS as a standalone filter ... This highlights LTS’s pivotal role in reducing portfolio volatility while maintaining favorable returns."

LTS is defined by subtracting κ·SVaR from LTA, and SVaR is itself a risk measure (Section 2.2.3). A fund passes the LTS filter only if LTS exceeds a threshold, so high-LTS funds are mechanically low-SVaR funds (for comparable LTA). The conclusion that LTS 'reduces portfolio volatility' is therefore largely a restatement of the feature's construction rather than an independently discovered property. The return level is empirical, but the risk-reduction claim is definitional.
fitted input called prediction [Section 2.3 and Section 3.4.3]
"Specifically, we set thresholds for each of the four features mentioned; a hedge fund must exceed all these thresholds to be considered likely to perform well in the next month. These thresholds are determined based on empirical analysis and historical market performance. ... Strategies that incorporate all available features for fund selection consistently outperform those that employ a more limited set of features or no feature-based filtering at all."

The thresholds are fitted to historical market performance, and the cumulative-return comparisons in Table 4 are computed over the same historical sample. The reported outperformance of comprehensive filter combinations is thus contaminated by in-sample threshold selection: the parameters are chosen after seeing the data and then evaluated on that same data. This is not an out-of-sample prediction of filter efficacy, so the claim that 'all available features' win is partly an artifact of the fitting procedure rather than a free-standing empirical result.

full rationale

The paper's central empirical comparisons (ML vs no-ML, equal vs AUM weighting, and the best-performer showcase) are generated by a moving-window XGBoost procedure and portfolio backtests; those are not derived from the features by construction and retain independent empirical content. I did not count the missing-value imputation in §3.2 as circularity: it is a serious validity threat, but it is a data-preprocessing assumption, not a step in which an output is equivalent to an input. Likewise, the citation to Cherny et al. (2010) for the PolyModel information-preservation theorem is a self-citation by co-author Douady, but it is a published theorem and is not the evidence for the empirical claims, so it does not by itself make the argument circular. The two circular components are (i) the LTS result, where the volatility-control finding is partly a tautology of the LTS = LTA − κ·SVaR definition, and (ii) the filter-comparison result, where thresholds described as 'determined based on empirical analysis and historical market performance' are then evaluated on the same historical performance, making the comprehensive-filter advantage partly an in-sample fit. These affect one of the three headline claims (PolyModel feature selection), while the ML and AUM-size comparisons are less directly affected, so the overall circularity is partial rather than total.

Assumptions & free parameters 8 free parameters · 5 assumptions · 0 invented entities

This ledger shows that the paper's central claims depend on many hand-chosen constants and unverified assumptions. The imputation of missing returns with -30 is both a free parameter and an unflagged assumption that can bias every downstream comparison. The filter thresholds, ridge penalty, risk-aversion parameter, and tail quantiles are not fully specified, limiting reproducibility.

free parameters (8)

Ridge penalty lambda = small positive number, not specified
Introduced to guarantee invertibility of the Hermite design matrix; a single small value is used for all regressions (Section 2.1.2).
LTS penalty kappa = 0.05
Set to 5% in the LTS definition LTS = LTA - kappa * SVaR (Section 2.2.6).
Risk aversion gamma in MRaR = Morningstar default 2; paper says it adjusts to risk flavor
No specific value is disclosed for the reported experiments (Section 2.2.1).
Filter thresholds for LTS, MRaR, Sharpe, and probability p_i = not disclosed
Chosen based on empirical analysis and historical market performance (Section 2.3); may include look-ahead.
Missing-value imputation constants = -30, -3, -1, -3
Arbitrary fill values for Return, Sharpe, LTS, and MRaR (Section 3.2).
Polynomial degree = 4
Chosen ad hoc to represent increasing, decreasing, convex, and concave curves (Section 2.1.1).
Tail quantiles and weights for LTA = 1%, 16%, 50%, 84%, 99% with Pareto tail fits
Used to compute long-term factor quantiles; weights are described as hyper-parameters (Section 2.2.4).
Number of target shuffles N = not specified
The permutation procedure for P-value scores requires N shuffles, but N is never stated (Section 2.1.3).

assumptions (5)

standard math Hermite polynomials form an orthogonal basis under the standard normal density
Invoked to justify invertibility of H_j H_j^T (Section 2.1.2).
domain assumption Risk factor log returns are approximately i.i.d. standard normal
Needed for orthogonality approximation (Section 2.1.2); not tested.
ad hoc to paper Degree-4 polynomial in one risk factor captures major relationships without overfitting
Assumed to balance flexibility and overfitting (Section 2.1.1).
domain assumption Target shuffling p-values identify causal relevance
Permutation test can detect spurious correlation but not causality; used for factor selection (Section 2.1.3).
ad hoc to paper Missing returns can be safely imputed with -30 and other constants
No justification; likely biases results if missingness is informative (Section 3.2).

how reviews work

0 comments

Cite this review

Pith. "Pith review of PolyModel for Hedge Funds' Portfolio Construction Using Machine Learning." pith.science (2026). https://pith.science/paper/QRUHYAP6

@misc{pith2026241211019,
  author       = {Pith},
  title        = {Pith review of: PolyModel for Hedge Funds' Portfolio Construction Using Machine Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QRUHYAP6}},
  note         = {Machine review of arXiv:2412.11019}
}

read the original abstract

The domain of hedge fund investments is undergoing significant transformation, influenced by the rapid expansion of data availability and the advancement of analytical technologies. This study explores the enhancement of hedge fund investment performance through the integration of machine learning techniques, the application of PolyModel feature selection, and the analysis of fund size. We address three critical questions: (1) the effect of machine learning on trading performance, (2) the role of PolyModel feature selection in fund selection and performance, and (3) the comparative reliability of larger versus smaller funds. Our findings offer compelling insights. We observe that while machine learning techniques enhance cumulative returns, they also increase annual volatility, indicating variability in performance. PolyModel feature selection proves to be a robust strategy, with approaches that utilize a comprehensive set of features for fund selection outperforming more selective methodologies. Notably, Long-Term Stability (LTS) effectively manages portfolio volatility while delivering favorable returns. Contrary to popular belief, our results suggest that larger funds do not consistently yield better investment outcomes, challenging the assumption of their inherent reliability. This research highlights the transformative impact of data-driven approaches in the hedge fund investment arena and provides valuable implications for investors and asset managers. By leveraging machine learning and PolyModel feature selection, investors can enhance portfolio optimization and reassess the dependability of larger funds, leading to more informed investment strategies.

Figures

Figures reproduced from arXiv: 2412.11019 by the authors.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Credit Risk Identification in Supply Chains Using Generative Adversarial Networks
cs.LG 2025-01 reject novelty 4.0 of 10

A GAN-based model is reported to beat baseline classifiers for supply chain credit risk, but the evaluation uses synthetic test data and no artifacts are provided.

Reference graph

Works this paper leans on

24 extracted references · 22 canonical work pages · cited by 1 Pith paper

[1]

Ali, A., X. Ma, S. Zawad, P. Aditya, I. E. Akkus, R. Chen, L. Yang, and F. Yan (2025). Enabling scalable and adaptive machine learning training via serverless computing on public cloud. Performance Evaluation\/ 167 , 102451

work page 2025
[2]

Cao, Z., S. Zhao, Z. Dong, and R. Douady (2024, September). Financial network analysis using polymodel theory. Preprints\/

work page 2024
[3]

Chen, T. and C. Guestrin (2016). Xgboost: A scalable tree boosting system. In Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining , pp.\ 785--794

2016
[4]

Chen, Z., F. Chen, L. Zhang, T. Ji, K. Fu, L. Zhao, F. Chen, L. Wu, C. Aggarwal, and C.-T. Lu (2020). Bridging the gap between spatial and spectral domains: A survey on graph neural networks. arXiv preprint arXiv:2002.11867\/

arXiv 2020
[5]

Douady, and S

Cherny, A., R. Douady, and S. Molchanov (2010). On measuring nonlinear risk with scarce observations. Finance and Stochastics\/ 14 , 375--395

work page 2010
[6]

The StressVaR: A New Risk Concept for Superior Fund Allocation

Coste, C., R. Douady, and I. I. Zovko (2009). The stressvar: A new risk concept for superior fund allocation. arXiv preprint arXiv:0911.4030\/

work page Pith review arXiv 2009
[7]

Dong, Z., X. Liu, B. Chen, P. Polak, and P. Zhang (2024). Musechat: A conversational music recommendation system for videos. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp.\ 12775--12785

work page 2024
[8]

Fung, W. and D. A. Hsieh (2004). Hedge fund benchmarks: A risk-based approach. Financial Analysts Journal\/ 60\/ (5), 65--80

work page 2004

Show all 24 references

[9]

Getmansky, M., P. A. Lee, and A. W. Lo (2015). Hedge funds: A dynamic industry in transition. Annual Review of Financial Economics\/ 7\/ (1), 483--577

2015
[10]

Wang, and I

Golbayani, P., D. Wang, and I. Florescu (2020). Application of deep neural networks to assess corporate credit rating. arXiv preprint arXiv:2003.02334\/

2020 arXiv
[11]

Guan, Y. (2019). PolyModel: Application in Risk Assessment and Portfolio Construction . Ph.\ D. thesis, Stony Brook University

2019
[12]

Zhu, and Y

Hu, T., W. Zhu, and Y. Yan (2023). Artificial intelligence aspect of transportation analysis using large scale systems. In Proceedings of the 2023 6th Artificial Intelligence and Cloud Computing Conference , pp.\ 54--59

2023
[13]

Kolanovic, M. and R. T. Krishnamachar (May 2017). Big data and ai strategies machine learning and alternative data approach to investing

2017
[14]

Ma, X., F. Yan, L. Yang, I. Foster, M. E. Papka, Z. Liu, and R. Kettimuthu (2024). Malletrain: Deep neural networks training on unfillable supercomputer nodes. In Proceedings of the 15th ACM/SPEC International Conference on Performance Engineering , pp.\ 190--200

2024
[15]

Han, and W

Ma, Y., R. Han, and W. Wang (2021). Portfolio optimization with return prediction using deep learning and machine learning. Expert Systems with Applications\/ 165 , 113973

2021
[16]

Chandola, V

Vijh, M., D. Chandola, V. A. Tikkiwal, and A. Kumar (2020). Stock closing price prediction using machine learning techniques. Procedia computer science\/ 167 , 599--606

2020
[17]

Wang, D. (2021). Application of Deep Learning to Corporate Credit Rating . Stevens Institute of Technology

2021
[18]

Wang, D., Z. Chen, I. Florescu, and B. Wen (2023). A sparsity algorithm for finding optimal counterfactual explanations: Application to corporate credit rating. Research in International Business and Finance\/ 64 , 101869

2023
[19]

Wang, and I

Wang, D., T. Wang, and I. Florescu (2020). Is image encoding beneficial for deep learning in finance? IEEE Internet of Things Journal\/ 9\/ (8), 5617--5628

2020
[20]

Wang, S., Y. Bai, K. Fu, L. Wang, C.-T. Lu, and T. Ji (2023). Alerta-net: A temporal distance-aware recurrent networks for stock movement and volatility prediction. arXiv preprint arXiv:2310.18706\/

2023 arXiv
[21]

Chen, and J

Xu, W., J. Chen, and J. Xiao (2024, November). A hybrid price forecasting model for the stock trading market based on ai technique

2024
[22]

Zhao, S. (2023). PolyModel: Portfolio Construction and Financial Network Analysis . Ph.\ D. thesis, Stony Brook University

2023
[23]

Zhao, S., Z. Dong, Z. Cao, and R. Douady (2024). Hedge fund portfolio construction using polymodel theory and itransformer

2024
[24]

Zhu, W. and T. Hu (2021). Twitter sentiment analysis of covid vaccines. In 2021 5th International Conference on Artificial Intelligence and Virtual Reality (AIVR) , pp.\ 118--122

2021

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [1]

Ali, A., X. Ma, S. Zawad, P. Aditya, I. E. Akkus, R. Chen, L. Yang, and F. Yan (2025). Enabling scalable and adaptive machine learning training via serverless computing on public cloud. Performance Evaluation\/ 167 , 102451

work page 2025

[2] [2]

Cao, Z., S. Zhao, Z. Dong, and R. Douady (2024, September). Financial network analysis using polymodel theory. Preprints\/

work page 2024

[3] [3]

Chen, T. and C. Guestrin (2016). Xgboost: A scalable tree boosting system. In Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining , pp.\ 785--794

2016

[4] [4]

Chen, Z., F. Chen, L. Zhang, T. Ji, K. Fu, L. Zhao, F. Chen, L. Wu, C. Aggarwal, and C.-T. Lu (2020). Bridging the gap between spatial and spectral domains: A survey on graph neural networks. arXiv preprint arXiv:2002.11867\/

arXiv 2020

[5] [5]

Douady, and S

Cherny, A., R. Douady, and S. Molchanov (2010). On measuring nonlinear risk with scarce observations. Finance and Stochastics\/ 14 , 375--395

work page 2010

[6] [6]

The StressVaR: A New Risk Concept for Superior Fund Allocation

Coste, C., R. Douady, and I. I. Zovko (2009). The stressvar: A new risk concept for superior fund allocation. arXiv preprint arXiv:0911.4030\/

work page Pith review arXiv 2009

[7] [7]

Dong, Z., X. Liu, B. Chen, P. Polak, and P. Zhang (2024). Musechat: A conversational music recommendation system for videos. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp.\ 12775--12785

work page 2024

[8] [8]

Fung, W. and D. A. Hsieh (2004). Hedge fund benchmarks: A risk-based approach. Financial Analysts Journal\/ 60\/ (5), 65--80

work page 2004