Publication Details

Abstract

Retail industries generate large amounts of transactional data that can be analyzed to predict future sales and optimize inventory management. Accurate sales prediction enables businesses to improve decision-making, reduce stock shortages, and enhance customer satisfaction. This research proposes a machine learning-based system for predicting product sales using synthetic retail data. The dataset includes product attributes such as item weight, item type, item price (MRP), outlet size, outlet location type, and outlet type. A regression-based machine learning model is developed using the scikit-learn implementation of the Gradient Boosting algorithm. The model predicts item outlet sales and evaluates performance using Mean Squared Error (MSE). Additionally, data visualization techniques and a graphical user interface dashboard are implemented to enhance user interaction and understanding of prediction results. Experimental results demonstrate that the proposed system can effectively predict retail sales and support business analytics.