Thematic and Dissemination

“Perfiles” (ISSN printed 1390-5740 - Electronic ISSN 2477-9105) is a scientific research journal published by the Science Faculty of “Escuela Superior Politécnica de Chimborazo” (Riobamba, Ecuador). It is being published following a biannual periodicity [(January - June) and (July- December)]. Its mission is to disseminate scientific-technological information related to predominant knowledge areas and related branches studied in the institution.

The main objective of the journal is the publication of original or review articles, short communications, technical reports, standards, specifications, letters to the editor, communications to congresses and, in short, other contents which are relevant to the scientific community.

“Perfiles” is a free access research journal. Its online version is available at the web address:
http://ceaa.espoch.edu.ec:8080/revista.perfiles/.
Investigation lines
  • Physical
  • Math
  • Chemistry
  • Biological Sciences
  • Health Sciences
  • Alternative energies


Article shipping format

Spanish Document English Document
Publication Instructions
Send your contributions to:
revistaperfiles@espoch.edu.ec



Issue Nº 27 Vol.1 - [January - June 2022]

CONTENT

AUTHORS: Jesús Eduardo Gamboa Unsihuay, Jesús Walter Salinas Flores

Abstract
The academic performance of a university student is generally measured through grades, which derive in a normal or deficient academic situation, depending in turn on several factors. The objective of this research was to find the main predictors of a university student's academic status after six semesters have elapsed since admission. For data analysis, the Boruta algorithm was used to select predictor variables and twelve classification algorithms were applied, after partitioning the data into training and evaluation sets. Then, those models with the best sensitivity, specificity and balanced accuracy values were chosen. Finally, an optimal assembly and cut-off point were used to improve predictions. The models with the best performance were logistic regression, Naive Bayes and vector support machines with linear kernel. When applying the optimal cut-off assembly, the specificity was 0.695 and sensitivity 0.947. The grade obtained in the mathematics course was one of the most important predictors of academic status after six semesters of study, while sociodemographic variables were not relevant.
Keywords: Ensemble, data mining, Boruta, optimal cut-off

Download:
                              

Panamerica Sur, Km 1 1/2- Escuela Superior Politécnica de Chimborazo - Science Faculty
Riobamba-Ecuador
Telephone: (593)(03) 2998200 ext.2221
Postal Code: EC060155
Email: revistaperfiles@espoch.edu.ec


        

    


    


        


    


    


    Creative Commons 4.0 License
   


    Code of ethics