Diffolio: A Diffusion Model for Multivariate Probabilistic Financial Time-Series Forecasting and Portfolio Construction
Authors:
- So-Yoon ChoSCSo-Yoon Cho
- Jin-Young KimJKJin-Young Kim
- Kayoung BanKBKayoung Ban
- Hyeng Keun KooHKHyeng Keun Koo
- Hyun-Gyoon KimHKHyun-Gyoon Kim
Preprint
Abstract
Probabilistic forecasting is crucial in multivariate financial time-series for constructing efficient portfolios that account for complex cross-sectional dependencies. In this paper, we propose Diffolio, a diffusion model designed for multivariate financial time-series forecasting and portfolio construction. Diffolio employs a denoising network with a hierarchical attention architecture, comprising both asset-level and market-level layers. Furthermore, to better reflect cross-sectional correlations, we introduce a correlation-guided regularizer informed by a stable estimate of the target correlation matrix. This structure effectively extracts salient features not only from historical returns but also from asset-specific and systematic covariates, significantly enhancing the performance of forecasts and portfolios. Experimental results on the daily excess returns of 12 industry portfolios show that Diffolio outperforms various probabilistic forecasting baselines in multivariate forecasting accuracy and portfolio performance. Moreover, in portfolio experiments, portfolios constructed from Diffolio's forecasts show consistently robust performance, thereby outperforming those from benchmarks by achieving higher Sharpe ratios for the mean-variance tangency portfolio and higher certainty equivalents for the growth-optimal portfolio. These results demonstrate the superiority of our proposed Diffolio in terms of not only statistical accuracy but also economic significance.Keywords
All Related Versions
- Version 1, 2025-11-10, ArXiv
- Version 2, 2026-03-29, ArXiv
- Published version: Information Fusion, 133.
This publication has 16 references indexed in Scilit:
- Generation of synthetic financial time series by diffusion modelsQuantitative Finance, 2025
- Fin-GAN: forecasting and classifying financial time series via generative adversarial networksQuantitative Finance, 2024
- Predicting Residential Energy Consumption by Explainable Deep Learning with Long-Term and Short-Term Latent VariablesCybernetics and Systems, 2022
- Explainable prediction of electric energy demand using a deep autoencoder with interpretable latent spaceExpert Systems with Applications, 2021
- Quant GANs: deep generation of financial time seriesQuantitative Finance, 2020
- Characteristics are covariances: A unified model of risk and returnJournal of Financial Economics, 2019
- Modeling financial time-series with generative adversarial networksPhysica A: Statistical Mechanics and its Applications, 2019
- Improved estimation of the covariance matrix of stock returns with an application to portfolio selectionJournal of Empirical Finance, 2003
- Industry costs of equityJournal of Financial Economics, 1997
- The Dividend-Price Ratio and Expectations of Future Dividends and Discount FactorsThe Review of Financial Studies, 1988