Prediction of E-Commerce Business Supplier's Purchase Quantity Based on LSTM-ARIMA Model and Kmeans Algorithm

With the rapid development of technology and market environment, e-commerce research is deepening and expanding. This paper takes the new technology marked by big data and artificial intelligence as the main axis, and utilizes the LSTM-ARIMA hybrid model to make real-time prediction of e-commerce merchants' incoming shipments and combines the Kmeans algorithm with Spearman indicators and so on, to improve the depth and breadth of the research. The research results of this paper show that (1) LSTM-ARIMA model is an important tool in the current and future e-commerce field to utilize big data and artificial intelligence technology, and it is an important way to improve the market competitiveness of e-commerce merchants. (2) In the research process, we focus on the optimization of the model and the configuration of the parameters, combining the AIC principle, Adam optimizer, etc., which improves the prediction accuracy of the model and avoids excessive prediction errors. (3) Due to the objectivity of the research, the model itself on the prediction data may lead to the existence of errors, the research conclusions and prediction accuracy are also the later research in this field need to be improved.