Predição do preço do leite utilizando dados de cooperativas.

Abstract

Currently, the volume of data generated is increasing more and more, making it possible to use techniques that detect patterns, generating relevant information for those who need it. One such technique is supervised learning, which is one of the machine learning techniques that deals with pattern recognition. This technique explores the construction of algorithms, capable of learning and making predictions from a set of data. Thus, a model trained with milk data would be able to suggest prices to be paid to the milk producer. Generally, the producer does not know for sure how much he will receive per liter of milk produced, as the milk is delivered over a month and he will only know the exact value in the next month, causing uncertainty. To alleviate this problem, this work trained a regression model capable of predicting the price of milk, obtaining metrics R² = 0.98 and RMSE = 0.014.