pith. machine review for the scientific record. sign in

arxiv: 2604.19476 · v2 · submitted 2026-04-21 · 💱 q-fin.PM · q-fin.GN· q-fin.ST

Recognition: unknown

Cross-Stock Predictability via LLM-Augmented Semantic Networks

Yikuan Huang , Zheqi Fan , Kaiqi Hu , Yifan Ye
Authors on Pith no claims yet

Pith reviewed 2026-05-10 01:14 UTC · model grok-4.3

classification 💱 q-fin.PM q-fin.GNq-fin.ST
keywords cross-stock predictabilityLLM edge filteringsemantic networks10-K embeddingsmean-reversion signalsfinancial graphseconomic relations
0
0 comments X

The pith

LLM filtering of edges in 10-K semantic networks raises the Sharpe ratio of cross-stock mean-reversion strategies from 0.742 to 0.820.

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

Text-based financial networks built from company disclosures frequently include spurious links that do not reflect actual economic ties, which weakens their value for predicting stock returns across firms. The paper constructs an initial sparse graph from similarities in 10-K filing embeddings and then applies a large language model to classify and retain only those candidate edges that represent genuine economic relations. The resulting refined graph aggregates pair-level mean-reversion signals into stock-level trading signals using weights that incorporate both relation type and distance. Backtests on S&P 500 constituents from 2011 to 2019 show that this LLM-filtered network delivers higher risk-adjusted returns and lower drawdowns than the unfiltered version. The results indicate that language-model reasoning can increase the economic accuracy of text-derived graphs used for cross-stock forecasting.

Core claim

The paper shows that a two-stage process—first forming a candidate graph from 10-K embeddings and then using an LLM to filter edges according to their economic content—produces a network that, when used to weight and combine mean-reversion signals, yields a long-short portfolio with a Sharpe ratio of 0.820 and maximum drawdown of -7.85 percent, compared with 0.742 and -10.47 percent for the unfiltered baseline on S&P 500 stocks between 2011 and 2019.

What carries the argument

Two-stage LLM-augmented semantic network: sparse candidate graph from embedding similarities followed by LLM classification of economic relations, then relation-aware and distance-based weighting of aggregated signals.

If this is right

  • LLM edge filtering reduces noise from non-economic links and thereby strengthens the aggregated trading signals.
  • Relation-aware weighting of mean-reversion pairs produces more accurate stock-level forecasts than uniform aggregation.
  • The refined graph structure can be reused for other text-based predictability tasks beyond mean reversion.
  • Performance gains appear in both higher returns per unit of risk and lower tail losses during the sample period.

Where Pith is reading between the lines

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

  • The approach could support periodic network updates as new 10-K filings become available without rebuilding the entire graph.
  • Similar LLM filtering might be tested on other disclosure types such as earnings transcripts to check for broader applicability.
  • The method offers a way to combine embedding-based scaling with targeted human-like judgment on link validity.

Load-bearing premise

The large language model correctly and consistently identifies which textual similarities reflect real economic connections rather than introducing bias or inconsistency.

What would settle it

A replication in which the LLM-filtered network produces no improvement or a deterioration in Sharpe ratio and drawdown when tested on the same 2011-2019 S&P 500 sample with varied prompts or models.

Figures

Figures reproduced from arXiv: 2604.19476 by Kaiqi Hu, Yifan Ye, Yikuan Huang, Zheqi Fan.

Figure 1
Figure 1. Figure 1: Overview of the proposed framework. Prompt Design. The LLM is instructed to act as an industry analyst and classify the relationship be￾tween two firms based only on the provided disclo￾sures. The model must select exactly one label from a predefined taxonomy: {competitor, supply chain, complementary, substitute, peer, unrelated}. The output is required in structured JSON format, including both the predict… view at source ↗
Figure 2
Figure 2. Figure 2: Cumulative returns of long-short portfolio and quintile sorting portfolios. [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Prompt template used for LLM-based edge classification. [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
read the original abstract

Text-based financial networks are increasingly used to study cross-stock return predictability. A common approach constructs links from similarities in firms' disclosure embeddings, but such networks often contain spurious edges because textual proximity does not necessarily imply economic connection. We propose a two-stage framework that first builds a sparse candidate graph from 10-K embeddings and then uses a large language model to classify and filter candidate edges according to their economic relations. The refined graph is used to aggregate pair-level mean-reversion signals into stock-level trading signals with relation-aware and distance-based weights. In a backtest on S&P 500 constituents from 2011 to 2019, LLM-based edge filtering improves the long-short Sharpe ratio from 0.742 to 0.820 and reduces maximum drawdown from $-$10.47% to $-$7.85%. These results suggest that LLM-based reasoning can improve the economic fidelity of text-derived financial networks and strengthen cross-stock predictability.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript proposes a two-stage framework for text-based financial networks: first constructing a sparse candidate graph from similarities in 10-K filing embeddings, then using a large language model to classify and filter candidate edges according to their economic relations. The refined graph aggregates pair-level mean-reversion signals into stock-level trading signals via relation-aware and distance-based weights. Backtests on S&P 500 constituents (2011–2019) report that LLM filtering raises the long-short Sharpe ratio from 0.742 to 0.820 and lowers maximum drawdown from −10.47% to −7.85%.

Significance. If the reported gains can be shown to arise from the LLM’s detection of economically meaningful relations rather than from any reduction in graph density, the work would provide a concrete method for improving the fidelity of embedding-derived networks in quantitative finance. The approach is timely given the rapid adoption of LLMs for semantic reasoning, and the empirical numbers are presented clearly enough to invite replication.

major comments (2)
  1. [Empirical results] Empirical results section: the headline comparison (Sharpe 0.742 vs. 0.820) pits the LLM-filtered graph only against the unfiltered sparse candidate graph. No ablation is reported that holds edge density fixed while changing the selection rule (e.g., random subsampling or non-LLM heuristics). Without this control, the performance lift cannot be attributed specifically to LLM economic classification rather than to the mechanical effect of pruning noisy pairs.
  2. [Methodology] Methodology section: the central assumption that the LLM accurately and consistently classifies edges as economically meaningful is not supported by any reported validation (human evaluation of classification accuracy, inter-annotator agreement, or tests for prompt sensitivity and hallucination). This validation is load-bearing for the claim that the refined graph improves economic fidelity.
minor comments (2)
  1. The abstract and results report concrete performance numbers but supply no information on the statistical significance of the Sharpe improvement or on robustness to alternative backtest windows.
  2. Details on the exact LLM prompt template, temperature setting, and any post-processing of LLM outputs are absent, making the procedure difficult to reproduce.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed and insightful comments on our manuscript. We agree with the concerns raised regarding the need for additional controls in the empirical analysis and validation of the LLM component. Below we provide point-by-point responses and indicate the revisions we plan to implement.

read point-by-point responses

  1. Referee: Empirical results section: the headline comparison (Sharpe 0.742 vs. 0.820) pits the LLM-filtered graph only against the unfiltered sparse candidate graph. No ablation is reported that holds edge density fixed while changing the selection rule (e.g., random subsampling or non-LLM heuristics). Without this control, the performance lift cannot be attributed specifically to LLM economic classification rather than to the mechanical effect of pruning noisy pairs.

    Authors: We acknowledge that the current comparison does not isolate the LLM's contribution from the effect of reduced graph density. In the revised version, we will include an ablation study that holds the number of edges constant by randomly subsampling the candidate graph to the same density as the LLM-filtered graph. We will report the Sharpe ratio and maximum drawdown for this random baseline alongside the LLM-filtered results. This will provide evidence on whether the performance gains stem specifically from the economic classification by the LLM. We will also explore a non-LLM heuristic such as keeping the top-k most similar pairs if it fits within the page limits. revision: yes

  2. Referee: Methodology section: the central assumption that the LLM accurately and consistently classifies edges as economically meaningful is not supported by any reported validation (human evaluation of classification accuracy, inter-annotator agreement, or tests for prompt sensitivity and hallucination). This validation is load-bearing for the claim that the refined graph improves economic fidelity.

    Authors: The referee is correct that we have not provided direct validation of the LLM's edge classifications in the submitted manuscript. To address this, we will add a section reporting a human evaluation study. Specifically, we will sample a subset of candidate edges, have them classified by the LLM, and then have domain experts (e.g., finance researchers) independently label whether the relation is economically meaningful. We will report accuracy, precision, recall, and inter-annotator agreement (Cohen's kappa). Additionally, we will test the sensitivity of results to different prompts and report any variations observed. These additions will substantiate the reliability of the filtering step. revision: yes

Circularity Check

0 steps flagged

No circularity in empirical backtest framework

full rationale

The paper describes a two-stage empirical procedure: construct a candidate graph from 10-K embeddings, apply LLM classification to filter edges, then aggregate mean-reversion signals into trading signals for a historical backtest on S&P 500 data 2011-2019. No equations, derivations, or first-principles claims are present that could reduce to self-definition, fitted parameters renamed as predictions, or self-citation chains. The reported Sharpe and drawdown figures are direct outputs of applying the procedure to the fixed historical returns; the LLM step operates on textual input independent of the performance metric. This is a standard empirical evaluation with no load-bearing circular steps.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Only the abstract is available, so the ledger is limited to the core assumption required by the filtering stage; no free parameters or invented entities are explicitly named.

axioms (1)
  • domain assumption Large language models can reliably identify economically meaningful relations from pairs of firm disclosures without systematic error or prompt dependence.
    The entire edge-filtering stage rests on this premise; no validation or error analysis is mentioned in the abstract.

pith-pipeline@v0.9.0 · 5471 in / 1274 out tokens · 33660 ms · 2026-05-10T01:14:11.129633+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

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

  1. From Hypotheses to Factors: Constrained LLM Agents in Cryptocurrency Markets

    q-fin.PM 2026-04 unverdicted novelty 7.0

    Constrained LLM agents discover cryptocurrency factors that produce a portfolio with 44.55% annualized return and Sharpe ratio of 1.55 in pure out-of-sample 2024-2026 testing after trading costs.

Reference graph

Works this paper leans on

181 extracted references · 4 canonical work pages · cited by 1 Pith paper · 1 internal anchor

  1. [1]

    Closing the

    Drost, Feike C and Werker, Bas JM , journal=. Closing the. 1996 , publisher=

    1996

  2. [2]

    Journal of Financial Economics , volume=

    Cross-stock momentum and factor momentum , author=. Journal of Financial Economics , volume=. 2023 , publisher=

    2023

  3. [3]

    Available at SSRN 5070808 , year=

    How Much of Cross-Stock Momentum Reflects Underreaction? , author=. Available at SSRN 5070808 , year=

  4. [4]

    Available at SSRN 4412580 , year=

    Can ChatGPT forecast stock price movements? Return predictability and large language models , author=. Available at SSRN 4412580 , year=

  5. [5]

    Available at SSRN 4416687 , year=

    Expected returns and large language models , author=. Available at SSRN 4416687 , year=

  6. [6]

    BloombergGPT: A Large Language Model for Finance

    BloombergGPT: A large language model for finance , author=. arXiv preprint arXiv:2303.17564 , year=

    work page internal anchor Pith review arXiv

  7. [7]

    From Text to Alpha: Can

    Choi, Chanyeol and Kim, Yoon and Yu, Yu and Cha, Young and Golkhou, V Zach and Halperin, Igor and Papaioannou, Georgios and Kim, Minkyu and Wang, Zhangyang and Kwon, Jihoon and others , journal=. From Text to Alpha: Can

  8. [8]

    Advances in Neural Information Processing Systems , volume=

    Finben: A holistic financial benchmark for large language models , author=. Advances in Neural Information Processing Systems , volume=

  9. [9]

    Structured Beliefs and Fund Performance: An

    Gao, Zhenyu and Xiong, Wei and Yuan, Jian , journal=. Structured Beliefs and Fund Performance: An

  10. [10]

    Pixiu: A large language model, instruction data and evaluation benchmark for finance.arXiv preprint arXiv:2306.05443,

    Pixiu: A large language model, instruction data and evaluation benchmark for finance , author=. arXiv preprint arXiv:2306.05443 , year=

    work page arXiv

  11. [11]

    A Financial Brain Scan of the

    Chen, Hui and Didisheim, Antoine and Somoza, Luciano and Tian, Hanqing , journal=. A Financial Brain Scan of the

  12. [12]

    Journal of Financial Economics , volume=

    Dual Peer Effects and Cross-Stock Predictability , author=. Journal of Financial Economics , volume=

  13. [13]

    On cross-stock predictability of peer return gaps in

    Chen, Yilin and Fan, Zheqi , journal=. On cross-stock predictability of peer return gaps in. 2026 , publisher=

    2026

  14. [14]

    The Journal of Finance , volume=

    Economic links and predictable returns , author=. The Journal of Finance , volume=. 2008 , publisher=

    2008

  15. [15]

    The Journal of Finance , volume=

    Do industries explain momentum? , author=. The Journal of Finance , volume=. 1999 , publisher=

    1999

  16. [16]

    The Review of Financial Studies , volume=

    Geographic lead-lag effects , author=. The Review of Financial Studies , volume=. 2020 , publisher=

    2020

  17. [17]

    Journal of Financial Economics , volume=

    Shared analyst coverage: Unifying momentum spillover effects , author=. Journal of Financial Economics , volume=. 2020 , publisher=

    2020

  18. [18]

    Management Science , volume=

    Empirical investigation of an equity pairs trading strategy , author=. Management Science , volume=. 2019 , publisher=

    2019

  19. [19]

    The Journal of Finance , volume=

    Principal portfolios , author=. The Journal of Finance , volume=. 2023 , publisher=

    2023

  20. [20]

    Journal of Financial and Quantitative Analysis , volume=

    Text-based industry momentum , author=. Journal of Financial and Quantitative Analysis , volume=. 2018 , publisher=

    2018

  21. [21]

    The Journal of Finance , volume=

    Market segmentation and cross-predictability of returns , author=. The Journal of Finance , volume=. 2010 , publisher=

    2010

  22. [22]

    Review of Financial Studies , volume=

    Jumps and stochastic volatility: Exchange rate processes implicit in deutsche mark options , author=. Review of Financial Studies , volume=. 1996 , publisher=

    1996

  23. [23]

    Quantitative Finance , volume=

    Pricing and calibration in the 4-factor path-dependent volatility model , author=. Quantitative Finance , volume=. 2025 , publisher=

    2025

  24. [24]

    Journal of Financial Economics , volume=

    The risk premia embedded in index options , author=. Journal of Financial Economics , volume=. 2015 , publisher=

    2015

  25. [25]

    Journal of Finance , volume=

    Model specification and risk premia: Evidence from futures options , author=. Journal of Finance , volume=. 2007 , publisher=

    2007

  26. [26]

    Handbook of Economic Forecasting , volume=

    Forecasting with option-implied information , author=. Handbook of Economic Forecasting , volume=. 2013 , publisher=

    2013

  27. [27]

    Journal of Econometrics , volume=

    Measuring tail risk , author=. Journal of Econometrics , volume=. 2024 , publisher=

    2024

  28. [28]

    Annual Review of Financial Economics , volume=

    Information in derivatives markets: Forecasting prices with prices , author=. Annual Review of Financial Economics , volume=

  29. [29]

    Journal of Economic Dynamics and Control , pages=

    Merton (1976) implied jump , author=. Journal of Economic Dynamics and Control , pages=. 2025 , volume=

    1976

  30. [30]

    Modeling the Implied Volatility Smirk in

    Ye, Yifan and Fan, Zheqi and Ruan, Xinfeng , journal=. Modeling the Implied Volatility Smirk in. 2025 , publisher=

    2025

  31. [31]

    Annual Review of Financial Economics , volume=

    Empirical option pricing models , author=. Annual Review of Financial Economics , volume=. 2022 , publisher=

    2022

  32. [32]

    Journal of Financial Economics , volume=

    Levy jump risk: Evidence from options and returns , author=. Journal of Financial Economics , volume=. 2014 , publisher=

    2014

  33. [33]

    Post-'87 crash fears in the

    Bates, David S , journal =. Post-'87 crash fears in the. 2000 , number =

    2000

  34. [34]

    Transform analysis and asset pricing for affine jump-diffusions , year =

    Duffie, Darrell and Pan, Jun and Singleton, Kenneth , journal =. Transform analysis and asset pricing for affine jump-diffusions , year =

  35. [35]

    Journal of Financial and Quantitative Analysis , volume=

    Fast filtering with large option panels: Implications for asset pricing , author=. Journal of Financial and Quantitative Analysis , volume=. 2024 , publisher=

    2024

  36. [36]

    A tale of two option markets: Pricing kernels and volatility risk , year =

    Song, Zhaogang and Xiu, Dacheng , journal =. A tale of two option markets: Pricing kernels and volatility risk , year =

  37. [37]

    Specification analysis of option pricing models based on time-changed

    Huang, Jing-Zhi and Wu, Liuren , journal=. Specification analysis of option pricing models based on time-changed. 2004 , publisher=

    2004

  38. [38]

    European Journal of Operational Research , volume=

    Smiles & smirks: Volatility and leverage by jumps , author=. European Journal of Operational Research , volume=. 2022 , publisher=

    2022

  39. [39]

    Assessing models of individual equity option prices , year =

    Bakshi, Gurdip and Cao, Charles and Zhong, Zhaodong , journal =. Assessing models of individual equity option prices , year =

  40. [40]

    Review of Financial Studies , volume=

    A closed-form solution for options with stochastic volatility with applications to bond and currency options , author=. Review of Financial Studies , volume=. 1993 , publisher=

    1993

  41. [41]

    Option valuation using the fast

    Carr, Peter and Madan, Dilip , journal=. Option valuation using the fast

  42. [42]

    A novel pricing method for

    Fang, Fang and Oosterlee, Cornelis W , journal=. A novel pricing method for

  43. [43]

    Efficient options pricing using the fast

    Kwok, Yue Kuen and Leung, Kwai Sun and Wong, Hoi Ying , booktitle=. Efficient options pricing using the fast. 2011 , publisher=

    2011

  44. [44]

    2003 , issn =

    The dynamics of stochastic volatility: Evidence from underlying and options markets , journal =. 2003 , issn =

    2003

  45. [45]

    2003 , issn =

    Alternative models for stock price dynamics , journal =. 2003 , issn =

    2003

  46. [46]

    Journal of Banking and Finance , volume =

    Volatility dynamics for the. Journal of Banking and Finance , volume =. 2012 , issn =

    2012

  47. [47]

    Journal of Business & Economic Statistics , volume=

    Empirical analysis of affine versus nonaffine variance specifications in jump-diffusion models for equity indices , author=. Journal of Business & Economic Statistics , volume=. 2015 , publisher=

    2015

  48. [48]

    Jump and volatility dynamics for the

    Yang, Hanxue and Kanniainen, Juho , journal=. Jump and volatility dynamics for the. 2017 , publisher=

    2017

  49. [49]

    Journal of Economic Dynamics and Control , volume =

    Model Complexity and Out-of-Sample Performance: Evidence from. Journal of Economic Dynamics and Control , volume =. 2018 , issn =

    2018

  50. [50]

    Journal of Finance , volume=

    Option profit and loss attribution and pricing: A new framework , author=. Journal of Finance , volume=. 2020 , publisher=

    2020

  51. [51]

    Interpretability in deep learning for finance: A case study for the

    Brigo, Damiano and Huang, Xiaoshan and Pallavicini, Andrea and de Oc. Interpretability in deep learning for finance: A case study for the. Risk Sciences , volume=. 2026 , publisher=

    2026

  52. [52]

    Chronologically consistent large language models.arXiv preprint arXiv:2502.21206,

    Chronologically consistent large language models , author=. arXiv preprint arXiv:2502.21206 , year=

    work page arXiv

  53. [53]

    Journal of Financial and Quantitative Analysis , volume=

    Machine learning and the stock market , author=. Journal of Financial and Quantitative Analysis , volume=. 2023 , publisher=

    2023

  54. [54]

    Journal of Financial and Quantitative Analysis , volume=

    Is technical analysis in the foreign exchange market profitable? A genetic programming approach , author=. Journal of Financial and Quantitative Analysis , volume=. 1997 , publisher=

    1997

  55. [55]

    Quantitative Finance , volume=

    Technical analysis as a sentiment barometer and the cross-section of stock returns , author=. Quantitative Finance , volume=. 2023 , publisher=

    2023

  56. [56]

    Xu, Yongxin and Xuan, Yuhao and Zheng, Gaoping , journal=

  57. [57]

    Available at SSRN 6489158 , year=

    Deep surrogate for non-affine stochastic volatility option valuation models , author=. Available at SSRN 6489158 , year=

  58. [58]

    Available at SSRN 4802345 , year=

    Trading Volume Alpha , author=. Available at SSRN 4802345 , year=

  59. [59]

    Available at SSRN 5974814 , year=

    On Options-Driven Realized Volatility Forecasting: Information Gains via Rough Volatility Model , author=. Available at SSRN 5974814 , year=

  60. [60]

    Journal of Portfolio Management , volume=

    Large language models for financial and investment management: Applications and benchmarks , author=. Journal of Portfolio Management , volume=. 2024 , publisher=

    2024

  61. [61]

    , author=

    Large Language Models for Financial and Investment Management: Models, Opportunities, and Challenges. , author=. Journal of Portfolio Management , volume=

  62. [62]

    Econometrica , volume=

    A simple, positive semi-definite, heteroskedasticity and autocorrelationconsistent covariance matrix , author=. Econometrica , volume=

  63. [63]

    Journal of Financial Markets , pages=

    Technical indicators and the cross-section of corporate bond returns in a machine learning era , author=. Journal of Financial Markets , pages=. 2025 , publisher=

    2025

  64. [64]

    The Journal of Finance , volume=

    Presidential address: The scientific outlook in financial economics , author=. The Journal of Finance , volume=. 2017 , publisher=

    2017

  65. [65]

    Chen, Jian and Tang, Guohao and Zhou, Guofu and Zhu, Wu , journal=

  66. [66]

    Generative

    Chai, Bailin and Jiang, Fuwei and Meng, Lingchao and You, Tian and Zhou, Guofu , journal=. Generative

  67. [67]

    European Journal of Operational Research , volume=

    Option valuation under no-arbitrage constraints with neural networks , author=. European Journal of Operational Research , volume=. 2021 , publisher=

    2021

  68. [68]

    Journal of Finance , volume=

    A nonparametric approach to pricing and hedging derivative securities via learning networks , author=. Journal of Finance , volume=. 1994 , publisher=

    1994

  69. [69]

    2016 , publisher=

    Deep Learning , author=. 2016 , publisher=

    2016

  70. [70]

    Journal of Financial Econometrics , volume=

    Forecasting value-at-risk using deep neural network quantile regression , author=. Journal of Financial Econometrics , volume=. 2024 , publisher=

    2024

  71. [71]

    Inferring volatility dynamics and risk premia from the

    Bardgett, Chris and Gourier, Elise and Leippold, Markus , journal=. Inferring volatility dynamics and risk premia from the. 2019 , publisher=

    2019

  72. [72]

    Journal of Financial Data Science , volume=

    Explainable machine learning models of consumer credit risk , author=. Journal of Financial Data Science , volume=

  73. [73]

    Journal of Portfolio Management , volume=

    Increasing the transparency of pricing dynamics in the US commercial real estate market with interpretable machine learning algorithms , author=. Journal of Portfolio Management , volume=

  74. [74]

    Explainable

    Misheva, Branka Hadji and Osterrieder, Joerg and Hirsa, Ali and Kulkarni, Onkar and Lin, Stephen Fung , journal=. Explainable

  75. [75]

    Journal of FinTech , volume=

    On deep calibration of (rough) stochastic volatility models , author=. Journal of FinTech , volume=. 2025 , publisher=

    2025

  76. [76]

    Journal of Financial Data Science , volume=

    Deep Calibration with Artificial Neural Network: A Performance Comparison on Option-Pricing Models , author=. Journal of Financial Data Science , volume=. 2023 , publisher=

    2023

  77. [77]

    Available at SSRN 4607397 , year=

    A new closed-form discrete-time option pricing model with stochastic volatility , author=. Available at SSRN 4607397 , year=

  78. [78]

    Available at SSRN , volume=

    Pricing and calibration in the 4-factor path-dependent volatility model , author=. Available at SSRN , volume=

  79. [79]

    Annals of Operations Research , pages=

    Forecasting gold price with the XGBoost algorithm and SHAP interaction values , author=. Annals of Operations Research , pages=. 2021 , publisher=

    2021

  80. [80]

    Review of Quantitative Finance and Accounting , pages=

    Bankruptcy prediction using machine learning and Shapley additive explanations , author=. Review of Quantitative Finance and Accounting , pages=. 2023 , publisher=

    2023

Showing first 80 references.