Diversification Across Predictors: Heterogeneous Ensemble Learning for Cross-Sectional Return Prediction

Paper: arXiv 2611.12389

Authors: Danning Li

Abstract

Cross-sectional stock return prediction remains challenging due to extremely low signal-to-noise ratios, high-dimensional factor spaces, and non-stationary predictor effectiveness. Individual model classes exhibit complementary failure modes in this regime: neural networks overfit noise due to high variance, while linear models underfit nonlinear structure due to high bias. We propose a heterogeneous ensemble framework that extends the financial principle of diversification to the predictor level. Through a bagging procedure combining linear predictors with nonlinear predictors, each model functions as a distinct specialist whose prediction errors exhibit low cross-covariance, enabling error cancellation upon aggregation. We formalize this mechanism via bias-variance-covariance decomposition, establishing an exact mathematical correspondence between portfolio diversification and predictor aggregation. On CSI300 daily-frequency Alpha158 factor datasets, the proposed ensemble achieves a Rank IC improvement exceeding 13% over standalone neural networks and surpasses tree-based state-of-the-art methods including XGBoost and Double Ensemble, as measured by both predictive accuracy and portfolio-based evaluation metrics.

Complexity vs Empirical Score

• Math Complexity: 7.5/10
• Empirical Rigor: 6.0/10 (capped from 8.0: transaction costs not modelled; see penalty flags)
• Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper presents a novel heterogeneous ensemble learning framework with a strong mathematical foundation in bias-variance-covariance decomposition. It demonstrates robust empirical performance on a relevant financial dataset, showcasing both predictive accuracy and portfolio-based evaluation metrics. The methodology is clearly articulated, suggesting good reproducibility.

Research Flowchart

  flowchart TD
    A[Research Goal: Improve Cross-Sectional Stock Return Prediction] --> B{Problem: Low Signal/Noise, High-Dim, Non-Stationary Predictors};
    B --> C[Methodology: Heterogeneous Ensemble Learning (Bagging)];
    C --> D{Core Idea: Diversification Across Predictors};
    D --> E[Formalization: Bias-Variance-Covariance Decomposition];
    E --> F[Data: CSI300 Daily Alpha158 Factors];
    F --> G[Computation: Model Training & Evaluation (Rank IC, Portfolio Metrics)];
    G --> H[Key Finding: Ensemble Achieves >13% Rank IC Improvement over NN, Outperforms XGBoost/Double Ensemble];