Predicting Future Sales: A Machine Learning Algorithm Showdown

In the internet era, handling vast data volumes manually is impractical, while accurate sales prediction remains crucial for organizations. Machine learning techniques offer powerful tools to extract hidden patterns from extensive datasets, enhancing prediction accuracy. This paper uses machine learning models to forecast future sales based on historical data from the “Store Item Demand Forecasting” dataset, comprising five years of sales data for 50 items across ten stores. Regression techniques, including linear regression, Random Forest regressor, and XGBoost, were employed, along with the LSTM algorithm. Results evaluated using MAE, RMSE, and R-squared, indicate that the XGBoost model outperformed other models in predicting sales with higher accuracy, closely followed by linear regression.