Trabalho completo:

Decisions in the industry are increasingly becoming data driven and the technological advance in communication provides huge data generation and availability, but only after being filtered and processed it can be transformed in useful information aiming to build knowledge and drive to intelligible conclusions. Machine learning algorithms are developed to this finality, these algorithms have wide applications range coupling a great power of information processing and prediction generation based on patterns. The automotive sector needs predictability and can take relevant advantages on future tendency predictions. This work presents an application of machine learning algorithm to estimate the energy efficiency of vehicles, expressed in MJ/km, to understand the potential application of some components in new vehicles. Some supervised machine learning algorithms are applied to predict the energy efficiency of a vehicle, based on constructive characteristics, such as mass, power, volumetric displacement, transmission, among others, aiming to obtain the highest possible reliability. Python programming interface was used, together with a database of passenger and light commercial vehicles, and comparisons were made among these algorithms aiming at this search with better performance.