Modelo basado en minería de datos para determinar factores de deserción estudiantil en la facultad de ciencias de la ingeniería y aplicadas de la universidad técnica de Cotopaxi.
Attrition in universities is considered as the abandonment of students that may be temporary or permanent, causing negative effects such as socioeconomic problems for the student and the institution that hosted him. For this reason, this problem must be treated in depth to establish strategies that...
Salvato in:
| Autore principale: | |
|---|---|
| Altri autori: | |
| Natura: | bachelorThesis |
| Lingua: | spa |
| Pubblicazione: |
2019
|
| Soggetti: | |
| Accesso online: | http://repositorio.utc.edu.ec/handle/27000/5328 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
Aggiungi Tag | Riassunto: | Attrition in universities is considered as the abandonment of students that may be temporary or permanent, causing negative effects such as socioeconomic problems for the student and the institution that hosted him. For this reason, this problem must be treated in depth to establish strategies that minimize the drop-out rates in universities and the successful completion of university studies. The literature review allowed to determine studies to predict attrition through data mining techniques, however, the studies analyzed were executed in Higher Education environments that differ from the context of education that is applied in Ecuador. A model is proposed to determine student dropout factors through data mining techniques to determine factors that influence dropout and its predictive influence. The experimental process is based on an online survey applied to 1457 students of the Faculty of Engineering and Applied Sciences of Engineering Careers: Electrical, Information Systems, Electromechanical and Industrial. The methodology applied corresponds to Knowledge Discovery in Databases (KDD), which consists of five stages: selection, pre-processing, transformation, extraction and interpretation and evaluation. The results found allow determining that the factors: inappropriate behavior in the classroom, bullying, teacher-student motivation, limited knowledge of the subject, addiction of social networks, emotional state, knowledge acquired in leveling, academic training, system of University entrance, family problems, residence and expectations regarding the selected career, are the factors that have the greatest influence on the drop-out of students in the Faculty of Engineering and Applied Sciences. While the J48, Random Forest and Sequential Minimal Optimization (SMO) data mining techniques resulted in a 92% dropout prediction rate. It is concluded that the use of data mining techniques can be considered as important for studies of the causes that affect students during their university student stay. In addition, this tool could be considered as a support tool for university authorities in order to establish strategies and policies to mitigate dropout rates. |
|---|