ESTUDO DE MODELOS DE CLASSIFICAÇÃO DE EVASÃO DAS UNIVERSIDADES UTILIZANDO TÉCNICAS DE APRENDIZADO DE MÁQUINA
Carregando...
Arquivos
Citações na Scopus
Tipo de produção
Trabalho de Conclusão de Curso
Data
2022-06-14
Autores
CAIO KOITI KANASHIRO
LUCAS CARVALHO DE OLIVEIRA SILVA
MATHEUS DE NARDI OLIVEIRA
LUCAS CARVALHO DE OLIVEIRA SILVA
MATHEUS DE NARDI OLIVEIRA
Orientador
Plinio Thomas Aquino Junior
Periódico
Título da Revista
ISSN da Revista
Título de Volume
Citação
Texto completo (DOI)
Palavras-chave
Resumo
Com o aumento perceptivo no nível de desistência de cursos do ensino superior, a causa
é algo que as faculdades procuram saber para tentar prever e assim evitar a evasão, para que,
não se tornem grandes problemas futuramente, tanto para o aluno como para o mercado de
trabalho, como, por exemplo, um grande arrependimento para o aluno que tomou uma decisão
precipitada e também causando grandes impactos no mercado de trabalho, tendo uma diminuição
de candidatos qualificados para exercer a profissão. Para isso foi desenvolvido um sistema
baseado de Machine Learning para prever a evasão de alunos e assim evitá-las. O sistema foi
proposto utilizando uma base de dados de universidades e aplicando o algoritmo K-Means,
onde foi possível perceber padrões de alunos desistentes, facilitando com que a faculdade tome
medidas para evitar a desistência do aluno. A metodologia aqui proposta foi capaz de mostrar
com uma taxa de mais de 90% de assertividade que atributos como idade são os que mais
interferem na evasão, já a área cursada, contrariamente, é o que menos interfere, demonstrando
assim que foi proposto neste trabalho uma metodologia efetiva para o controle da evasão escolar,
no auxílio da gestão institucional
With the evident increase in student drop out level of higher education courses, univer- sities are researching the possible causes of student drop out. By understanding the reasons students leave and avoiding big issues in the future, whether it’s snap decisions students make that can lead to regret, or causing the job market to decrease the number of qualified applicants to the professions, universities can attempt to predict future resignations and hopefully avoid them. Due to that, a Machine Learning-based system was developed to predict abandonment and thus avoid it. The system is presented using data from universities and applying the K- Means algorithm, making it possible to identify patterns related to student drop out and thus helping the university take preemptive action to avoid future resignations. The method propo- sed here was capable of showing, with over 90% of accuracy, that attributes such as age is the one that interfere the most relating to resignations. The course itself, on the other hand, is the one that interferes the least. These findings indicate that what is proposed in this assignment is an effective method to control university drop out numbers along with the support of the institutional management.
With the evident increase in student drop out level of higher education courses, univer- sities are researching the possible causes of student drop out. By understanding the reasons students leave and avoiding big issues in the future, whether it’s snap decisions students make that can lead to regret, or causing the job market to decrease the number of qualified applicants to the professions, universities can attempt to predict future resignations and hopefully avoid them. Due to that, a Machine Learning-based system was developed to predict abandonment and thus avoid it. The system is presented using data from universities and applying the K- Means algorithm, making it possible to identify patterns related to student drop out and thus helping the university take preemptive action to avoid future resignations. The method propo- sed here was capable of showing, with over 90% of accuracy, that attributes such as age is the one that interfere the most relating to resignations. The course itself, on the other hand, is the one that interferes the least. These findings indicate that what is proposed in this assignment is an effective method to control university drop out numbers along with the support of the institutional management.