Paper: arXiv 2605.01176

Authors: Yi Wang, Takashi Hasuike

Abstract

Decision-focused learning (DFL) is attractive for portfolio optimization because it trains predictors according to downstream decision quality rather than prediction accuracy alone. However, SPO(Smart, Predict then Optimize surrogate)-based DFL may produce inflated return signals and unstable portfolio reallocations. This study provides a KKT-based interpretation showing that portfolio decisions can be viewed as ranking over risk- and transaction-cost-adjusted marginal scores. Empirically, we examine prediction inflation and excessive turnover in SPO-trained portfolios, and evaluate clipping, min-max rescaling, and partial portfolio adjustment as practical stabilization mechanisms. The results suggest that realistic output constraints and portfolio-level turnover control improve the implementability of SPO-based portfolio strategies.

Complexity vs Empirical Score

  • Math Complexity: 6.5/10
  • Empirical Rigor: 5.5/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper combines formal KKT optimality condition derivations and Lagrangian analysis with empirical experiments evaluating stabilization techniques like clipping and rescaling within a portfolio optimization framework.

Research Flowchart

  flowchart TD
    A[<b>Goal:</b> Analyze prediction inflation & <br/>instability in SPO-based Portfolio Optimization] --> B[<b>Data/Inputs:</b> Financial Market Data <br/>& Equities Portfolios]
    B --> C[<b>Methodology:</b> KKT-based Interpretation <br/>of SPO Ranking Mechanisms]
    C --> D[<b>Processes:</b> Simulate SPO-training <br/>& Test Stabilization Mechanisms]
    D --> E[<b>Mechanisms:</b> Clipping, Rescaling, <br/>& Partial Adjustment]
    E --> F[<b>Outcomes:</b> Realistic constraints & turnover <br/>control improve strategy implementability]

Cited References (17)

  1. Return Prediction for Mean-Variance Portfolio Selection: How Decision-Focused Learning Shapes Forecasting Models (2025) — link
  2. Decision-Focused Learning: Through the Lens of Learning to Rank (2022)
  3. Portfolio Selection (1952)
  4. Prediction-based portfolio optimization model using neural networks (2009)
  5. Decision-making for financial trading: A fusion approach of machine learning and portfolio selection (2019)
  6. Portfolio optimization with return prediction using deep learning and machine learning (2021)
  7. Trading on the Edge: Neural, Genetic, and Fuzzy Systems for Chaotic Financial Markets (1994)
  8. Deep neural networks, gradient-boosted trees, random forests: Statistical arbitrage on the S&P 500 (2017)
and 9 more
  1. Smart ``Predict, then Optimize’’ (2022)
  2. Decision-focused learning: Foundations, state of the art, benchmark and future opportunities (2024)
  3. Optimizing mean conditional value-at-risk portfolios through deep neural network stock prediction (2025) — link
  4. Melding the data-decisions pipeline: Decision-focused learning for combinatorial optimization (2019)
  5. Integrating prediction in mean-variance portfolio optimization (2023) — link
  6. Distributionally robust end-to-end portfolio construction (2023) — link
  7. End-to-end, decision-based, cardinality-constrained portfolio optimization (2025) — link
  8. Estimating Covariance for Global Minimum Variance Portfolio: A Decision-Focused Learning Approach (2025) — link
  9. Semi-Decision-Focused Learning with Deep Ensembles: A Practical Framework for Robust Portfolio Optimization (2025) — link

0 of 17 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.