Research on Demand Assessment and Forecasting of Express Service Based on TOPSIS and LSTM Models

With the rapid development of Internet technology and the e-commerce market, online shopping has become an important part of daily life, and accurate prediction of the demand for courier services is crucial for courier companies to optimize their warehouses and transportation networks. In this paper, a comprehensive model of evaluation, prediction, and optimization is constructed for express demand and transportation problems. Firstly, the important cities were screened out by preprocessing the data and analyzing them in four dimensions: the receiving volume, the shipping volume, the trend of changes in the number of expresses and the correlation with other cities, calculating the weights of the indicators using the entropy method, and conducting a comprehensive evaluation using the TOPSIS method. Secondly, the time-series changes of express delivery volume were counted, and the identification of periodicity was optimized by establishing the SARIMA model, which improved the prediction accuracy of the model, and the value of R ²reached 0.86. Finally, the LSTM neural network was successfully used to predict the route cargo volume. This study provides an effective data analysis and forecasting tool for the express service industry.