Integrated GM-LSTM-KAN Model for Forecasting NEV Market Dynamics

This study combines the Grey Prediction Model GM(1,1) with an integrated model of Long Short-Term Memory (LSTM) and Knowledge-Aware Network (KAN) to forecast the sales trends of New Energy Vehicles (NEVs) in China and analyze their impact on carbon dioxide emissions. The GM(1,1) model is employed for short-term forecasting of NEV sales, while LSTM is utilized for capturing long-term temporal dependencies in carbon emission data. Additionally, KAN incorporates NEV-related covariates to enhance the model’s predictive accuracy. The results indicate that the promotion of NEVs significantly slows the growth of carbon emissions, providing valuable data support for China’s carbon peak and carbon neutrality goals. This study offers scientific guidance and decision-making support for policymakers in formulating carbon reduction strategies and NEV promotion policies.