Research on Cargo Volume Forecasting Based on the ARIMA-LSTM Hybrid Model

This paper proposes a cargo volume forecasting method based on the ARIMA-LSTM hybrid model, aiming to improve the accuracy and reliability of predictions. We collected historical data from 57 sorting centers, covering a time span from day 0 to 121, to establish and verify the forecasting model. After data preprocessing, including filling in missing values and handling outliers, the data quality was ensured. On this basis, we constructed a hybrid ARIMA-LSTM model that combines the linear feature capturing ability of the ARIMA model with the non-linear relationship processing capability of the LSTM model, to more comprehensively capture the characteristics and patterns in the data. Through experiments, we verified the performance of the model in forecasting at different logistics centers, finding that the ARIMA model performs well when dealing with data with obvious seasonal and periodic characteristics, while the LSTM model has advantages in handling complex nonlinear relationships and long-term dependencies. The experimental results show that the hybrid model can effectively predict the cargo volume for each sorting center over the next 30 days, providing reliable predictive outcomes.