S3G: A Stock State-Space Graph for Financial Time-Series Modeling
International Journal of Business and Social Science Studies · 2026
DOI: 10.66795/ijbsss32 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Business and Social Science Studies, 2026. Verfasst von Noah Laurent. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Financial time-series modeling presents significant challenges due to the highly stochastic, non-stationary, and interconnected nature of global equity markets. Traditional deep learning architectures predominantly process individual stock trajectories in isolation or rely on static correlation matrices that fail to capture the dynamic, latent relationships between assets over time. To address these fundamental limitations, this paper introduces the Stock State-Space Graph framework, an innovative architectural paradigm that synergizes the temporal expressiveness of state-space models with the relational learning capabilities of dynamic graph neural networks. By formulating each financial asset as a continuous-time state-space node and dynamically inferring the evolving edge connections through state transitions, the proposed framework captures both complex intra-series temporal dependencies and inter-series spatial correlations. Comprehensive empirical evaluations conducted on large-scale, real-world financial datasets demonstrate that the proposed architecture significantly outperforms contemporary baseline methodologies, including recurrent neural networks, temporal convolutional networks, and static graph-based approaches. The results indicate marked improvements in predictive accuracy, robustness to market volatility, and computational efficiency, thereby establishing a new state-of-the-art foundation for multifaceted financial forecasting and portfolio optimization tasks.
Zitieren
APA
Laurent, N. (2026). S3G: A Stock State-Space Graph for Financial Time-Series Modeling. International Journal of Business and Social Science Studies. https://doi.org/10.66795/ijbsss32
Vancouver
Laurent N. S3G: A Stock State-Space Graph for Financial Time-Series Modeling. International Journal of Business and Social Science Studies. 2026. doi:10.66795/ijbsss32
BibTeX
@article{Laurent2026,
author = {Laurent, Noah},
title = {S3G: A Stock State-Space Graph for Financial Time-Series Modeling},
journal = {International Journal of Business and Social Science Studies},
year = {2026},
doi = {10.66795/ijbsss32}
}
EUVIMED