A Stacked LSTM Model to Predict CO2 Emissions in Vietnam: Evidence From International Economic Factors for International Tourism, FDI, and Export Dependence

Vietnam’s rapid economic growth has intensified environmental pressures, particularly from international tourism, foreign direct investment (FDI), and export-oriented production. This study investigates how these sectors collectively influence national CO 2 emissions. Using data from 1990 to 2023 sourced from the Vietnam General Statistics Office, and World Bank, this study first applies quantitative analysis to assess the correlations between these drivers and CO 2 levels. An advanced Stacked Long Short-Term Memory (LSTM) model, with optimized preprocessing, is then used to forecast emissions from 2024 to 2030. The model achieves high predictive accuracy (RMSE = 0.1201, R 2 = .8657) and consistently outperforms traditional methods such as ARIMA, KNN, Naive Bayes, and standard neural networks. Results reveal strong positive links between CO 2 emissions and FDI, exports, and tourism, with projections indicating continued growth in emissions, potentially stabilizing after 2027. These findings offer valuable insights for developing policies that balance economic growth with environmental sustainability.