Today, the achievement of the graduate profile competencies is very important evidence that shows the quality of university programs. However, universities must wait for their students to finish or reach the end of their studies in order to accurately measure their achievement, which brings problems that when some do not reach the expected levels there is not enough time to take action corrective. In this sense, in the literature, there are several investigations related to the graduate profile and others that focused on the prediction of academic performance in face-to-face or virtual courses, in the prediction of student desertion, in predicting the student's academic motivation, in predicting the placement of a student in a job, in predicting the final global average of the university student or predicting the delay in graduation, but there are no studies that have directly addressed the subject of the prediction of the level of achievement of the profile of the university graduate engineering. Due to this, the general objective of the research was to develop a computational model based on Machine Learning that allows early warning of the level of achievement of the graduation profile of engineering university students, in such a way that timely information is available to make corrective decisions in advance. and not wait until the end of the studies to obtain this result. For this, the CRISP-DM methodology was followed, data was collected from 1982 graduates of different engineering programs of a Peruvian university and using Matlab and Machine Learning algorithms, in the end a fairly accurate model was created (accuracy: 96.7% ) using the SVM algorithm and it was found that the characteristics related to the grades obtained up to the IV cycle of studies in Mathematics and Physics courses are the best predictors to predict the level of achievement of the graduation profile of engineering students.

Hoy en día el logro de las competencias del perfil de egreso es una evidencia muy importante que muestra la calidad de los programas universitarios. Sin embargo, las universidades deben esperar a que sus estudiantes terminen o lleguen al final de sus estudios para poder medir exactamente el logro de éstas, lo cual trae problemas que cuando algunos no alcancen los niveles esperados ya no haya el tiempo suficiente para poder tomar acciones correctivas. En ese sentido, en la literatura, existen varias investigaciones relacionadas con el perfil de egreso y otras que se centraron en la predicción del rendimiento académico en cursos presenciales o virtuales, en la predicción de la deserción estudiantil, en predecir la motivación académica del estudiante, en predecir la colocación de un estudiante en un empleo, en predecir el promedio final global del estudiante universitario o predecir el retraso en la graduación, más no hay estudios que hayan abordado directamente el tema de la predicción del nivel de logro del perfil del egresado universitario de ingeniería. Debido a esto, la investigación tuvo como objetivo general desarrollar un modelo computacional basado en Machine Learning que permita alertar tempranamente el nivel de logro del perfil de egreso de los estudiantes universitarios de Ingeniería, de tal manera que se tenga información oportuna para tomar decisiones correctivas anticipadamente y no esperar hasta el final de los estudios para poder obtener este resultado. Para esto, se siguió la metodología CRISP-DM, se recolectaron datos de 1982 egresados de diferentes programas de ingeniería de una universidad peruana y haciendo uso de Matlab y algoritmos de Machine Learning, al final se creo un modelo bastante preciso (accuracy: 96.7%) utilizando el algoritmo SVM y se obtuvo que las características relacionadas con las calificaciones obtenidas hasta el IV ciclo de estudios en los cursos de Matemáticas y Física, son los mejores predictores para pronosticar el nivel de logro del perfil de egreso de los estudiantes de ingeniería.

COMPUTATIONAL MODEL TO ALERT EARLY THE LEVEL OF ACHIEVEMENT OF THE GRADUATION PROFILE OF UNIVERSITY ENGINEERING STUDENTS / ZELADA VALDIVIESO, HÉCTOR MIGUEL. - (2022 Sep 30).

COMPUTATIONAL MODEL TO ALERT EARLY THE LEVEL OF ACHIEVEMENT OF THE GRADUATION PROFILE OF UNIVERSITY ENGINEERING STUDENTS

ZELADA VALDIVIESO, HÉCTOR MIGUEL
2022-09-30

Abstract

Today, the achievement of the graduate profile competencies is very important evidence that shows the quality of university programs. However, universities must wait for their students to finish or reach the end of their studies in order to accurately measure their achievement, which brings problems that when some do not reach the expected levels there is not enough time to take action corrective. In this sense, in the literature, there are several investigations related to the graduate profile and others that focused on the prediction of academic performance in face-to-face or virtual courses, in the prediction of student desertion, in predicting the student's academic motivation, in predicting the placement of a student in a job, in predicting the final global average of the university student or predicting the delay in graduation, but there are no studies that have directly addressed the subject of the prediction of the level of achievement of the profile of the university graduate engineering. Due to this, the general objective of the research was to develop a computational model based on Machine Learning that allows early warning of the level of achievement of the graduation profile of engineering university students, in such a way that timely information is available to make corrective decisions in advance. and not wait until the end of the studies to obtain this result. For this, the CRISP-DM methodology was followed, data was collected from 1982 graduates of different engineering programs of a Peruvian university and using Matlab and Machine Learning algorithms, in the end a fairly accurate model was created (accuracy: 96.7% ) using the SVM algorithm and it was found that the characteristics related to the grades obtained up to the IV cycle of studies in Mathematics and Physics courses are the best predictors to predict the level of achievement of the graduation profile of engineering students.
30-set-2022
ACCARDO, AGOSTINO
33
2019/2020
Settore ING-INF/06 - Bioingegneria Elettronica e Informatica
Università degli Studi di Trieste
File in questo prodotto:
File Dimensione Formato  
_Tesis_hzelada_V02.pdf

accesso aperto

Descrizione: Tesis Final
Tipologia: Tesi di dottorato
Dimensione 2.05 MB
Formato Adobe PDF
2.05 MB Adobe PDF Visualizza/Apri
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11368/3030921
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact