Research on inventory demand forecasting based on K-means and BP neural network

• 1、School of Management, Wuhan University of Science and Technology, Wuhan, 430065

Abstract：Due to the high uncertainty and volatility of market demand, enterprises are facing the heavy responsibility of cost reduction and inventory optimisation, so they have certain requirements for the accuracy of inventory demand planning, and with the increase of uncertainty in the external environment and the growth of supply chain risk, the difficulty of inventory demand forecasting for enterprises is also getting bigger and bigger. In this regard, this study selects the monthly inventory quantity of some materials of Company Y, a vehicle-mounted intelligent terminal enterprise, from 2022 to 2024 as the object of analysis, takes the characteristics of material demand as the index, adopts K-means clustering algorithm to classify the materials, and uses the BP neural network algorithm to predict the inventory quantity of the next three months for each class of materials. The results show that using K-means and BP neural network forecasting model for inventory demand prediction, its predicted value is closer to the real value compared with the enterprise\'s prediction, and the accuracy of the model prediction is relatively high. Therefore, compared with the enterprise\'s current forecasting method, the accuracy is higher, and it has strong applicability and feasibility.

Keywords： BP neural network K-means clustering inventory demand forecasting material requirements forecasting model