Machine Learning Based Retail Profit Prediction for Business Decision Support – A Comparative Model Analysis

Abstract: Retail organizations generate large volumes of transactional data, making accurate profit estimation essential for effective business decision-making. This study proposes a machine learning-based framework for predicting order- level retail profit using the SuperStore Orders dataset containing 51,290 records and 21 attributes. Three regression models—Linear Regression, Decision Tree Regressor, and Random Forest Regressor—were developed and evaluated using Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and Coefficient of Determination (R²). Experimental results demonstrate that the Random Forest Regressor achieved the highest predictive performance with an R² score of 0.7095, outperforming the other models by effectively capturing complex non-linear relationships among retail transaction features. The findings indicate that Random Forest is a reliable approach for retail profit prediction and can serve as a valuable predictive component in Business Decision Support Systems.

Keywords: Machine Learning, Random Forest, Retail Profit Prediction, Business Decision Support System, Regression Analysis