The automotive recall data search and its analysis applying machine learning

Abstract

Paper aims: This article investigates the worldwide trend of growth in the number of recalls, as well as in the number of products involved in each campaign. Originality:To investigate these facts, a study of the automotive recall was developed, comprising Brazil, the European Union, and the United States of America.Research method: Due to the different availabilities between the locations, search tools and software were developed to obtain and group hidden data from 2010 to 2019. Main findings:In this work, the impacts of the recall were analyzed using three categories of algorithms: clustering, classification, and regression. Analyzes were made about the results obtained and discussions were built about the importance of applying the machine learning technique. Implications for theory and practice:The use of search tools and software to obtain and group hidden data in databases and opens the opportunity for new research in various areas.