Year: 2021
Type: article
Abstract: Predicting the individual performance of each student can provide valuable information as to which students are at greatest risk of failure or dropout, and consequently highlight which characteristics negatively influence the student's academic life.Data mining provides the tools necessary to address this educational data in the search for knowledge and patterns that can be obtained.Therefore, this work uses an educational database obtained at the UCI machine learning repository related to students grades in Portuguese and proposes models using extreme learning machine networks, ensemble learning and optimization by particle swarm in order to predict students' grades.In addition two simulated data sets were also used to verify the consistency of the results obtained through the proposed regression models.After obtaining the error value for each proposed model, hypothesis tests were performed to ascertain the veracity of the results.The results indicate a better performance of the model that combines the ensemble learning, particle swarm optimization and extreme learning machine networks.
Institution: Universidade de Pernambuco
Language: English
Subfield: Computer Science Applications
Field: Computer Science
Domain: Physical Sciences
SDG: Quality education
Open Access status: diamond
Award: G29316538
OpenAlex ID: W3133054306