A Decision model for merchandise replenishment based on logistic regression and ARIMA time series forecasting

In supermarkets or shopping malls that mainly deal with fresh food, vegetable commodities are generally difficult to store for a long time with a short freshness period. And the quality of vegetables and storage time present a certain relationship, the longer the storage time the worse the quality of vegetables, and even the existence of varieties if the day did not sell the next day will not be able to sell. Therefore, supermarkets need to replenish goods on the same day according to the past sales situation and the demand of residents. To address the problem, this paper fits the data to derive a linear equation for sales volume and cost-plus pricing using a logistic regression model and Adam's optimized gradient descent algorithm. Finally, a time series ARIMA forecasting model is used to give daily replenishment and specific pricing for the week ahead based on past daily sales data. In the results, we can adjust the merchandise based on predicted replenishment and pricing strategies. The model reveals the supermarket's replenishment and pricing strategy for fresh food in general, and the supermarket can adjust its replenishment strategy based on its data.