USO DE MODELOS DE PREDIÇÃO PARA MELHORAR O PLANEJAMENTO DE ESTOQUE DE ALTA ROTAÇÃO - ESTUDO EM UMA EMPRESA NO SETOR DE BEBIDAS

Abstract

This project develops a prediction model for inventory planning of high-turnover products, specifically draft beer, in a company in the beverage sector. Using neural network and time series techniques, the aim is to minimize waste and guarantee product availability, taking into account the short shelf life of draft beer and the need for advance orders. The research involved collecting sales data from previous years, processing the data and developing a user-friendly interface. The model was trained with machine learning algorithms to predict weekly sales. The results show that the model is effective at predicting sales patterns, allowing for adjustments and future integrations, such as the inclusion of climate and demographic data. This work provides a solid basis for optimizing inventory management and formulating more efficient commercial strategies.