A Novel Forecasting Framework for the New Energy Vehicle Industry in the Global Trade Ecosystem

The rise of the New Energy Vehicle (NEV) industry has significantly influenced global trade flows, supply chains, and economic policies. However, traditional forecasting models—such as time-series techniques or regression-based neural networks struggle to effectively capture the complex, interconnected structure of global trade relationships. To surpass these limitations, this study proposes a novel forecasting framework based on Graph Neural Networks (GNNs), which are inherently suited for modelling relational and structured data. In the proposed model, countries are represented as nodes and their NEV-related trade connections as weighted edges, enabling the system to learn from both localized economic indicators and broader international interdependencies. The approach is trained on a curated dataset derived from Kaggle’s “Global Trade Analysis” repository, which provides detailed multi-country trade statistics and macroeconomic variables relevant to various sectors, including the NEV domain. Experimental results reveal that the GNN-based model surpasses traditional neural architectures in forecasting trade volumes and predicting economic shifts within the NEV sector. Specifically, the proposed model achieves an accuracy of 97.8%, a precision of 96.5%, a recall of 96.9%, an F1-score of 96.7%, and an AUC-ROC score of 0.982, thereby validating the efficacy of GNNs in modeling the real-world complexities of international economics. This approach presents a promising decision-support tool for policymakers and stakeholders to anticipate market dynamics and respond proactively in the fast-evolving landscape of sustainable transportation.