Prediction of the optimal demand for products in a food company.

This thesis comprehensively addresses the importance of demand forecasting in the food industry and reviews the existing literature in this field. Demand forecasting plays a strategic role in critical processes such as inventory management, logistics planning, and customer satisfaction. The sensitivity of food products to seasonal variations and uncertainties in consumer behavior makes this process more complex and delicate. In the literature review, the strengths and weaknesses of methods used for demand forecasting, particularly regression models, time series analysis, and machine learning approaches, have been thoroughly examined. Moreover, the dynamics specific to the food industry are highlighted, and the importance of determining effective modeling strategies for the sector is emphasized. The thesis focuses on the data collection, processing, and analysis processes for demand forecasting. Both internal data sourcesand external factors have been taken into account; different scenarios were analyzed using the publicly available Walmart dataset. In the data preprocessing stage, missing data was analyzed and outliers were detected, ensuring the data was suitable for the modeling process. These methods have improved the accuracy and reliability of the forecasting models. In the third section, the forecasting performance of different models, such as SARIMAX, Random Forest, Gradient Boosting, and LSTM, has been compared. Machine learning-based models were found to outperform traditional methods due to their ability to model nonlinear relationships and achieve higher accuracy. Methods such as Random Forest and Gradient Boosting were also effective in understanding data diversity and identifying feature importance. In conclusion, this thesis underscores the critical role of demand forecasting methods in making strategic decisions in the food industry; by comparing different modeling approaches, it provides a valuable contribution to both academic literature and industry practices.