Abstract
This study explores the application of Dynamic Bayesian Networks (DBNs) for modeling and forecasting the relationship between education financing and macroeconomic indicators, particularly Gross Domestic Product (GDP). The proposed DBN model is constructed through collaboration with domain experts and incorporates statistical data on funding volumes and institutional metrics across different educational levels. A static Bayesian network was initially developed and learned on real-world data from 2013 to 2023, achieving a prediction accuracy of 74.36%. This network was subsequently extended to a dynamic structure by integrating temporal dependencies between variables. The resulting DBN enables the simulation of various funding scenarios, allowing for adaptive forecasting of GDP based on changes in student numbers, institutional capacity, and investment levels. The model demonstrates the capacity to account for uncertainty, temporal variation, and incomplete data characteristics inherent in socioeconomic processes. The sensitivity analysis and experimental validation confirm the model's practical applicability in strategic decision-making for educational finance planning. The work highlights the potential of DBNs as a powerful tool for developing intelligent, evidence-based policy models in the context of digital transformation in education governance.