The Statistical Significance of the Inclusion of Graph Neural Networks in the Financial Time Series Forecasting Problem

2026-05-20

AI Summary

This work proposes the Time-Geometric model to address the limitation of traditional financial time series forecasting methods, which predominantly focus on temporal patterns while overlooking the potential value of geometric structural information. For the first time, this study systematically demonstrates the statistically significant contribution of geometric patterns in univariate financial time series to predictive performance. The proposed approach leverages graph neural networks to extract intrinsic geometric features from the data and integrates them into a hybrid architecture with classical time series models. Extensive experiments on multiple real-world financial datasets show that the Time-Geometric model substantially improves prediction accuracy, thereby validating the effectiveness and practical utility of incorporating geometric information into time series modeling.

Abstract

Forecasting univariate time series in the financial market is a challenging endeavor. While numerous statistical and machine learning models have been introduced to address this challenge, they typically concentrate solely on analyzing temporal patterns within the time series data. In this research, we study the statistical significance of the inclusion of geometric patterns in enhancing forecasting accuracy within the context of time series analysis. We introduce the Time-Geometric model, a combination of models designed to exploit both geometric and temporal patterns. The contribution of this research lies in advancing the domain of univariate time series prediction,as demonstrated through extensive empirical evaluations. Our findings underscore that leveraging geometric patterns, captured through Graph Neural Networks, yields statistically significant improvements in forecasting accuracy.

Problem

• Financial Time Series Forecasting
• Graph Neural Networks
• Geometric Patterns
• Statistical Significance
• Univariate Time Series

Innovation

• Graph Neural Networks
• Financial Time Series Forecasting
• Geometric Patterns
• Time-Geometric Model
• Statistical Significance