Optimization of Food Supply Chain Management Using Ma‎chine Learning

This research work aims to improve supply chain management using predictive modeling techniques. First, we gained a comprehensive ‎understanding of the company's operations, internal policies, product categories, customer base, and procurement processes. Through a ‎literature review, we identified machine learning algorithms, including multivariate linear regression, support vector regression (SVR), re-‎gression decision trees, and neural networks, and implemented these algorithms using libraries such as Scikit-learn and TensorFlow. After ‎data preparation and normalization, we trained and evaluated the models using various performance metrics such as R2, MAE, MAPE, ‎MSE, and RMSE. Among all the models, the SVR algorithm had the highest predictive performance. After applying the predictive model, ‎supply chain metrics improved: monthly revenue per supermarket decreased by 36.36%, monthly revenue per category decreased by ‎‎39.74%, and total revenue decreased by 25.95%. These findings confirm the effectiveness of machine learning in improving demand fore-‎casting and supply chain efficiency. Recommendations include creating a data-driven culture within the company, exploring other regression ‎algorithms, integrating historical data for new products, and continuously improving the predictive model. In addition, there is a need to ‎improve evaluation metrics and continuously monitor model performance to improve the process‎.