Protótipo de um sistema inteligente para a prevenção de acidentes graves em empresas

Abstract

In the work routine, accidents cause several losses, both for the contractor and the people hired. At the same time, profit and materials are lost in the companies, the work capacity and, in the worst cases, the worker’s life is affected. From 2002 to 2020, a rate of 6 deaths was recorded for every 100 thousand formal jobs in Brazil. The present work resulted in the evaluation of a computational model using machine learning to prevent serious accidents. Through the use of sliding window algorithms for data standardization, satisfactory results were achieved, in which it was possible to predict the occurrence of severe accidents for periods of up to seven days. Additionally, a prototype of an intelligent web system was proposed, using cloud microservices, using the prediction model created.