Regression-Based Demand Forecasting for Retail Supply Chain Management in Aba Commercial District

This study applies regression-based statistical methods to demand forecasting for retail supply chain management in the Aba Commercial District of Abia State, South East Nigeria, one of West Africa's most important centres for textile, footwear, and light manufacturing trade. Accurate demand forecasting is a foundational requirement for efficient inventory management, procurement planning, and logistics optimisation in retail supply chains, and the informal and semi-formal character of most Aba-based commercial enterprises has meant that quantitative forecasting methods are rarely employed. The study works with weekly sales records voluntarily provided by 25 wholesale and retail traders in the Ariaria International Market and the Shoe Market covering the period January 2022 to December 2023. Seasonal decomposition of each product category's sales time series is performed prior to model fitting. Multiple linear regression models with seasonal dummy variables, trend terms, and lagged sales as predictors are estimated and compared against univariate ARIMA and Holt-Winters exponential smoothing benchmarks in terms of out-of-sample forecast accuracy. Inventory optimisation calculations integrating the demand forecasts with economic order quantity and reorder point models are developed to demonstrate the operational value of the forecasting output. Results show that the regression model with seasonal dummies and a linear trend achieves a mean absolute percentage error of 8.7 percent, a 34 percent improvement over the naive benchmark used by traders in the survey. Keywords: demand forecasting, regression analysis, retail supply chain, inventory management, Aba