Enhancing Supermarket Inventory Management Through Arima and Sarimax - Based Demand Forecasting

Employing time series analysis, particularly the Auto Regressive Integrated Moving Average (ARIMA) model and its seasonal extension, Seasonal Auto Regressive Integrated Moving Average with Exogenous Variables (SARIMAX), the study aims to estimate demand for various retail commodities. The dataset includes past and present product sales information from a supermarket, spanning from 2015 to 2018, which is used to predict future demand. Accurate demand forecasting is essential for effective inventory management and strategic decision-making in the retail sector. By taking seasonality and external influences into account, SARIMAX improves the model's capacity to identify complex patterns and trends in time series data. Our project provides forecasts and evaluates their accuracy using Mean Squared Error (MSE) calculations and time series cross-validation. The primary dataset, which focuses on a variety of retail products, provides a wealth of historical insights. The initiative aims to deliver accurate demand forecasts to support inventory optimization and informed business decisions. The accuracy of these projections is rigorously assessed using various evaluation techniques, including time series cross-validation and MSE estimates. The model's ability to adapt to seasonal variations and external demand drivers is further enhanced by incorporating SARIMAX, making it a robust tool for retail demand forecasting.