Paper: arXiv 2508.20108

Authors: Hyunwoo Lee, Jihyeong Jeon, Jaemin Hong, U Kang

Abstract

How can we address distribution shifts in stock price data to improve stock price prediction accuracy? Stock price prediction has attracted attention from both academia and industry, driven by its potential to uncover complex market patterns and enhance decisionmaking. However, existing methods often fail to handle distribution shifts effectively, focusing on scaling or representation adaptation without fully addressing distributional discrepancies and shape misalignments between training and test data. We propose ReVol (Return-Volatility Normalization for Mitigating Distribution Shift in Stock Price Data), a robust method for stock price prediction that explicitly addresses the distribution shift problem. ReVol leverages three key strategies to mitigate these shifts: (1) normalizing price features to remove sample-specific characteristics, including return, volatility, and price scale, (2) employing an attention-based module to estimate these characteristics accurately, thereby reducing the influence of market anomalies, and (3) reintegrating the sample characteristics into the predictive process, restoring the traits lost during normalization. Additionally, ReVol combines geometric Brownian motion for long-term trend modeling with neural networks for short-term pattern recognition, unifying their complementary strengths. Extensive experiments on real-world datasets demonstrate that ReVol enhances the performance of the state-of-the-art backbone models in most cases, achieving an average improvement of more than 0.03 in IC and over 0.7 in SR across various settings.

Complexity vs Empirical Score

  • Math Complexity: 7.0/10
  • Empirical Rigor: 6.0/10 (capped from 8.5: transaction costs not modelled; see penalty flags)
  • Quadrant: Holy Grail — high math complexity, high empirical rigor (effective empirical rigor capped from 8.5)

Why this score: The paper includes advanced mathematical concepts like Geometric Brownian Motion and stochastic processes, alongside rigorous empirical validation with real-world datasets, code availability, and specific performance metrics (IC, SR).

Research Flowchart

  flowchart TD
  A["Research Goal: Mitigate distribution shifts<br>in stock price data to improve prediction"] --> B["Data Processing: Raw Stock Price Data"]
  B --> C["ReVol Methodology"]
  
  subgraph C ["ReVol Framework"]
      C1["Step 1: Normalization<br>Remove return, volatility, price scale"] --> C2["Step 2: Attention Module<br>Estimate characteristics to reduce anomalies"] --> C3["Step 3: Reintegration<br>Restore traits lost during normalization"]
  end
  
  C3 --> D["Hybrid Modeling"]
  
  subgraph D ["Computational Processes"]
      D1["Long-term: Geometric Brownian Motion"]
      D2["Short-term: Neural Networks"]
  end
  
  D1 & D2 --> E["Final Prediction"]
  
  E --> F["Key Findings"]
  
  subgraph F ["Experimental Outcomes"]
      F1["Avg IC improvement > 0.03"]
      F2["Avg SR improvement > 0.7"]
      F3["Enhances SOTA backbones"]
  end

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes ReVol, a Return‑Volatility Normalization method that explicitly addresses distribution shift in stock price prediction.

Source sentence
We propose ReVol (Return-Volatility Normalization for Mitigating Distribution Shift in Stock Price Data), a robust method for stock price prediction that explicitly addresses the distribution shift problem.

Assumes 1

  • Stock prices follow a geometric Brownian motion (GBM) model. model
    Source sentence
    We assume that stock prices follow a GBM, except for the independence assumption (refer to property 2 of GBM in Section 2.1).

Data

Backtest sample: US, China, UK, Korea stocks · equities · daily

  • stock market datasets market-prices unknown
    Source sentence
    We use stock market datasets collected from the stock markets of the United States, China, United Kingdom, and Korea.

Costs

Not modeled A trading or execution result is reported without costs.

Finds 2

Reported performance (ReVol): Sharpe 2.63

  • ReVol improves prediction performance, yielding on average >0.03 increase in IC and >0.7 increase in Sharpe Ratio over baseline models. performance
    Source sentence
    On average, ReVol improves IC by over 0.03 and SR by more than 0.7 compared to the baselines, confirming its ability to enhance predictive accuracy and profitability.
  • ReVol consistently outperforms other normalization methods (RevIN and Dish‑TS) across all markets in both IC and SR. performance
    Source sentence
    ReVol shows consistent improvement in IC and SR across all markets, outperforming the other normalization methods.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 57 related papers). Coverage is partial, so this is neither replication nor refutation.

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.

Cited References (30)

  1. Speaker normalization for self-supervised speech emotion recognition (2022)
  2. A regression approach to speech enhancement based on deep neural networks (2014)
  3. A dual-stage attention-based recurrent neural network for time series prediction (2017)
  4. Enhancing Stock Movement Prediction with Adversarial Training. (2019)
  5. Accurate multivariate stock movement prediction via data-axis transformer with multi-level contexts (2021)
  6. Arbitrage theory in continuous time (2019)
  7. Regime switching models (2010) — link
  8. Fractional Brownian motions, fractional noises and applications (1968)
and 22 more
  1. ARIMA models and the Box–Jenkins methodology (2011)
  2. Deep learning-based stock price prediction using LSTM and bi-directional LSTM model (2020)
  3. Stock market’s price movement prediction with LSTM neural networks (2017)
  4. Stock price prediction using attention-based multi-input LSTM (2018)
  5. MASTER: Market-Guided Stock Transformer for Stock Price Forecasting (2024)
  6. Modeling the stock relation with graph network for overnight stock movement prediction (2021)
  7. FreQuant: A Reinforcement-Learning based Adaptive Portfolio Optimization with Multi-frequency Decomposition (2024)
  8. Meta-adaptive stock movement prediction with two-stage representation learning (2024)
  9. Learning multiple stock trading patterns with temporal routing adaptor and optimal transport (2021)
  10. Adam: A method for stochastic optimization (2014) — link
  11. Dish-ts: a general paradigm for alleviating distribution shift in time series forecasting (2023)
  12. On information and sufficiency (1951)
  13. Doubleadapt: A meta-learning approach to incremental learning for stock trend forecasting — → 2306.09862
  14. Reversible Instance Normalization for Accurate Time-Series Forecasting against Distribution Shift (2021) — link
  15. Stochastic processes for physicists: understanding noisy systems (2010)
  16. Answering the skeptics: Yes, standard volatility models do provide accurate forecasts (1998)
  17. The distribution of realized stock return volatility (2001)
  18. Learning phrase representations using RNN encoder-decoder for statistical machine translation (2014) — link
  19. Attention is all you need (2017)
  20. Attention-based autoregression for accurate and efficient multivariate time series forecasting (2021)
  21. Accurate stock movement prediction with self-supervised learning from sparse noisy tweets (2022)
  22. Accurate Stock Movement Prediction via Multi-Scale and Multi-Domain Modeling (2024)

1 of 30 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.