Aplicación del machine learning para la clasificación y predicción e importación de repuestos tipo crítico para automotores.

Abstract

This research aimed to implement a Machine Learning model for the classification and prediction of critical spare parts imports in the automotive sector. The problem is based on the reliance on traditional inventory management methods in small and medium-sized enterprises, which leads to supply chain inefficiencies, increased operational costs, and difficulties in demand forecasting. Methodologically, an unsupervised learning model, K-Means clustering, was applied to classify spare parts according to their level of criticality (high, medium, and low), considering variables such as turnover rate, monetary impact, calculated stock, and inventory inconsistency. Subsequently, a predictive model using Random Forest Regressor was developed to estimate the purchases variable, and its performance was evaluated using Mean Absolute Error (MAE) and the coefficient of determination (R²). The results showed a high concentration of spare parts classified as critical and strong predictive performance, with an MAE of 1.18 units and an R² of 0.943, indicating high explanatory capacity. It is concluded that the application of Machine Learning techniques significantly improves inventory management, optimizes import planning, and strengthens business competitiveness in the automotive sector.