• Victor Elias Palasios Silva Universidade Evangélica de Goiás - UniEVANGÉLICA
  • José Victor Rocha Silvestre Universidade Evangélica de Goiás - UniEVANGÉLICA
  • Denis Rodrigues de Faria Universidade Evangélica de Goiás - UniEVANGÉLICA
  • Pedro Moreira de Oliveira Neto Universidade Evangélica de Goiás - UniEVANGÉLICA
  • Luana Lucieli Aguiar Freitas Universidade Evangélica de Goiás - UniEVANGÉLICA
  • Raul Antônio Vagmacker Ribeiro Universidade Evangélica de Goiás - UniEVANGÉLICA


Education, Evasion, College Education, Management Indicators, Software, Machine Learning


With the advancement of the field of science and technology, the extreme relationship of
dependence between man and this area, made the human being doomed to always seek a
new technology. One of the areas that suffered impacts generated by this constant evolution
was education, which needs to constantly adapt in terms of technology. At a given moment,
both elementary and technical education underwent moments of advances induced by this
technological growth. Because of this, college education, one of those affected, is hit with a
problem that plagues most institutions, which is the dropout of students during the current
period, due to the volatility of scientific progress. In order to alleviate this recurrence, the work aims to build a tool for the generation of reports, which can help in the understanding and reduction of the impacts that the dropout of students can cause. To initiate the project, a
bibliographic research was carried out in the areas of education, evasion and machine
learning to better understand the scope of the problem and the areas that should be focused
on to reach the best conclusion. At the same time, artifacts were developed to increase the
veracity of the solution, including PM Canva and the production methodology for organization, personas, the Product Backlog Building to prove requirements, in addition to the test plan, architecture and structuring of the database. data, so that the demonstration of the possibility and viability of the product generated during the work was palpable, so that it was possible to introduce the implementation of the algorithm on the real database and, thus, accurately present the possibilities of evasions through such reports.