Quantifying and Forecasting the Impact of Governmental Internet Service Capability on the Digital Economy: A Machine Learning-Integrated Panel Data Approach

The digital economy has become a key influencing factor of global economic growth, and government internet service capacity has gotten increasing attention because of its role in facilitating digital transformation. Using panel data of nine cities in China's Greater Bay Area from 2018 to 2022, this study empirically examines and predicts the effects of government Internet service capacity on the development of the digital economy. This paper also uses regression models and machine learning architectures-long-term short-term memory (LSTM) and recurrent neural networks (RNN). There are five key findings. Government internet service capacity has a statistically significant positive impact. Control variable analysis shows that gross regional product (GRP) and education spending significantly contribute to digital economy growth. Local general budget revenues and the number of industrial enterprises above a certain size have negative impacts. Predictive modeling shows a consistent upward trend in the policy effect of government Internet service capacity after estimation by the model of the LSTM and RNN. Heterogeneity analysis across cities revealed significant differences, with first-tier cities such as Guangzhou (LSTM, 2023, 3.3232) and Shenzhen showing higher predicted impacts compared to Zhaoqing (RNN, 2024, 0.9728) and Huizhou. The empirical validation of the reinforcing effect of digital governance improvement on economic digitization confirms the sustainable policy effectiveness. This study offers the link between the interdisciplinary linkages between public administration and the digital economy. Furthermore, this study also offers evidence-based strategies for optimizing government digital service frameworks and fostering high-quality digital economic ecosystems.