Research on E-Commerce Sales Forecasting Based on ARIMA

E-commerce sales forecasting is crucial for optimising inventory management and improving supply chain efficiency. In this paper, a hybrid forecasting model combining K-means clustering and ARIMA is proposed to solve the multiwarehouse commodity demand forecasting problem. Firstly, the commodities are clustered into 75 classes by shipping characteristics through the improved K-means algorithm, and an ARIMA model is constructed based on the clustering results to predict the regular commodity demand (1- wmape =0.5852). Empirical evidence shows that the hybrid model is better than a single method in terms of stability and accuracy, providing theoretical support for e-commerce intelligent warehousing.