Inventory prediction based on CNN-LSTM Model

Abstract

Demand forecasting and inventory optimization are the core issues of e-commerce supply chain management, which are crucial for ensuring timely delivery of goods, reducing inventory costs, and improving inventory turnover efficiency. This paper is based on Python's merge function to merge table data and transcode text data. Then, by drawing a q-q diagram and K-S test to determine the distribution of the data columns, it was found that the shipment volume data follows a normal distribution. Adopting 3σ The principle is to determine outliers and manually determine that the two data with larger values in the edge values are considered outliers. Replace the above outliers with missing values and use Newton interpolation for linear filling to obtain the standard dataset after data preprocessing. We establish a prediction model based on attention mechanism for prediction, and ultimately obtain more accurate prediction results.