Enhancing Black-Litterman Portfolio via Hybrid Forecasting Model Combining Multivariate Decomposition and Noise Reduction

2025-05-03

AI Summary

To address the bottleneck in the Black-Litterman (BL) model wherein subjective views rely on low-accuracy asset price forecasts, this paper proposes a hybrid SSA-MA-EMD-TCN forecasting framework. First, singular spectrum analysis (SSA) is synergistically integrated with multivariate-aligned empirical mode decomposition (MA-EMD) to suppress financial time-series noise and decouple multi-scale features; subsequently, a temporal convolutional network (TCN) is end-to-end integrated for joint multi-asset forecasting. This approach significantly improves forecast accuracy and enhances the BL model’s capability for dynamic view construction. In out-of-sample short-horizon portfolio experiments on a 20-stock NASDAQ-100 subset, the proposed method achieves a 12.7% higher annualized return, a 19.3% lower volatility, and a 31.5% improvement in Sharpe ratio compared to mean-variance optimization, equal-weighting, and market-cap-weighting benchmarks.

Technology Category

• Machine Learning: Multi-instance/Multi-view Learning
• Search and Optimization: Mixed Discrete/Continuous Search
• Reasoning under Uncertainty: Stochastic Optimization

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

The sensitivity to input parameters and lack of flexibility limits the traditional Mean-Variance model. In contrast, the Black-Litterman model has attracted widespread attention by integrating market equilibrium returns with investors' subjective views. This paper proposes a novel hybrid deep learning model combining Singular Spectrum analysis (SSA), Multivariate Aligned Empirical Mode Decomposition (MA-EMD), and Temporal Convolutional Networks (TCNs), aiming to improve the prediction accuracy of asset prices and thus enhance the ability of the Black-Litterman model to generate subjective views. Experimental results show that noise reduction pre-processing can improve the model's accuracy, and the prediction performance of the proposed model is significantly better than that of three multivariate decomposition benchmark models. We construct an investment portfolio by using 20 representative stocks from the NASDAQ 100 index. By combining the hybrid forecasting model with the Black-Litterman model, the generated investment portfolio exhibits better returns and risk control capabilities than the Mean-Variance, Equal-Weighted, and Market-Weighted models in the short holding period.

Problem

• Improving Black-Litterman model's subjective view accuracy via hybrid forecasting
• Addressing sensitivity and flexibility issues in Mean-Variance portfolio models
• Enhancing portfolio returns and risk control with noise-reduced predictions

Innovation

• Hybrid deep learning with SSA and MA-EMD
• Noise reduction improves prediction accuracy
• TCNs enhance Black-Litterman subjective views