Modelo preditivo de vendas para uma franquia de Bubble Tea para recomendação de estoque

Abstract

Organizations are constantly seeking to ensure sustainability through supply and demand. Inventories represent significant components and must be managed in such a way that financial resources are used in the most rational way possible. Inefficient inventory management causes unwanted effects like capital immobilization and increased inventory maintenance costs. Even with the advancement of computational technologies, not all Brazilian companies use formal quantitative methods for inventory management, especially small and medium-sized ones, although some companies use it to deal with administrative routines, they can use data related to consumption of products to forecast demands and provision your inventory. The research aimed to analyze the consumption pattern of a bubble tea franchise and applied prediction techniques comparing the results generated. The machine learning technique based on Generalized Linear Models was the one that presented the best result, contributing to the most adequate provisioning of stock to meet demand, according to the consumption pattern.