Machine learning applications in the areas of production engineering

Abstract

This paper carried out a systematic review and a bibliometric analysis to assess quantitatively and qualitatively the papers that applied machine learning (ML) techniques published between 2015 and 2020 on the main Brazilian scientific events related to Production Engineering (PE). A research protocol along with a set of eligibility criteria were devised to identify research that contained applications of ML techniques, resulting in 71 articles being selected for the bibliometric analysis. This study measured the number of papers per event, publications per institution, most frequent authors, cooperation networks, most explored PE areas, as well as the main ML tasks and techniques being used. The results show that Quality Management and Economic Engineering are the PE areas in which most ML applications take place, while ML tasks aimed at classification and algorithms based on Neural Networks and Decision Trees were the most frequent in the selected papers. As for the main research institutions involved with this research, the Federal University of Technology – Paraná (UTFPR) was the most active in the period analyzed, and along with the Pontifical Catholic University of Paraná (PUC-PR) and the Fluminense Federal University, these were the three top universities that most participated in research that involved partnership with other universities. The researcher who published the most participated in six papers and belongs to the Graduate Program in Production and Systems Engineering (PPGEPS) at PUC-PR.