STORE SALES – TIME SERIES FORCASTING

Abstract: Accurate sales forecasting plays a crucial role in the retail industry by enabling effective inventory management, resource allocation, and demand planning. This study presents a hybrid time series forecasting approach to predict daily store sales by combining statistical and machine learning models. The proposed method integrates Linear Regression, Ridge Regression, and Facebook Prophet models to capture both linear and nonlinear dependencies in the sales data. Historical store sales records are preprocessed by handling missing values and extracting time-based features such as day, month, year, and day of the week. The ensemble prediction is obtained by averaging the outputs of all three models. Experimental results demonstrate that the ensemble model achieves low Mean Absolute Error (MAE) and Root Mean Square Error (RMSE), effectively capturing seasonal variations and trends in store-level sales data.

Keywords: Time series forecasting, sales prediction, Linear Regression, Ridge Regression, Prophet, ensemble model.