ProFish-LSTM: A Hybrid Deep Learning Model for Accurate Economic Growth and Inflation Forecasting

With the continuous development of China’s economy, accurately forecasting economic growth (GDP) and inflation (CPI) trends is of significant importance for policymaking, market investment and international trade. However, traditional economic forecasting methods, such as ARIMA and SARIMA, often perform inadequately when dealing with nonlinear fluctuations and sudden events. To address this issue, this paper proposes a hybrid deep learning model, ProFish-LSTM, based on the Prophet model, multilayer LSTM and artificial fish swarm algorithm (AFSA), aimed at improving the accuracy of forecasting China’s economic growth and inflation. Through experiments on four public datasets, the proposed model demonstrates excellent performance across various metrics. The mean squared error (MSE) is reduced by approximately 30% compared with the traditional ARIMA model, the mean absolute error (MAE) decreases by 25% and the coefficient of determination (R 2 ) improves by 0.15. This indicates that the model excels in both long-term trend forecasting and short-term fluctuation capture, outperforming traditional ARIMA, SARIMA models and standalone LSTM models. Furthermore, ProFish-LSTM outperforms existing hybrid models on all datasets, showcasing its strong predictive accuracy and stability. This research provides a new approach to economic forecasting, enhancing the accuracy and stability of economic time series prediction through the combination of deep learning and optimization algorithms, with significant application potential and theoretical value.