Predicting Agricultural Product Unit Production Using the K-Nearest Neighbors Algorithm

This paper presents a study focused on the analysis of phytosanitary production. The objective is to predict pro-duction per unit, in order to optimize production efficiency, as an essential key in the supply chain. The data for this study encompasses production data, product names, and input param-eters (electricity usage, water consumption, and raw materials), with production as the target variable. The machine learning technique applied for forecasting is the K-Nearest Neighbors regression (KNN-R) algorithm, known for its robust performance, especially on small datasets. This results in a predictive model with a mean absolute error (MAE) of 0.052 and a mean squared error (MSE) of 0.0057. This research provides valuable insights for agricultural and food companies seeking to improve their production forecasting methods and enhance the efficiency of agricultural product manufacturing processes.