Modelo para la determinación de factores de deserción estudiantil en la Universidad Técnica de Cotopaxi utilizando Minería de Datos
Higher education in Ecuador plays a very important role in the search for development and social welfare, thus becoming a main axis for the national development. The interest in approaching the investigation to determine dropout factors at the Technical University of Cotopaxi, is due to the fact tha...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis de maestría |
| Estado: | Versión publicada |
| Fecha de publicación: | 2020 |
| País: | Ecuador |
| Institución: | Universidad Técnica de Cotopaxi |
| Repositorio: | Repositorio Universidad Técnica de Cotopaxi |
| Idioma: | español |
| OAI Identifier: | oai:oai:repositorio.utc.edu.ec:27000:27000/7143 |
| Acceso en línea: | http://repositorio.utc.edu.ec/handle/27000/7143 |
| Access Level: | acceso abierto |
| Palabra clave: | DESERCIÓN ESTUDIANTIL FACTORES DESERCIÓN MINERÍA DE DATOS REDES NEURONALES INFORMÁTICA Y SISTEMAS COMPUTACIONALES |
| Sumario: | Higher education in Ecuador plays a very important role in the search for development and social welfare, thus becoming a main axis for the national development. The interest in approaching the investigation to determine dropout factors at the Technical University of Cotopaxi, is due to the fact that university student dropout has become a current problem that affects the student, their family environment, universities and society in general. Therefore, the determination of dropout factors in universities can be considered as a key strategy for decision making institutional. For the development of the research, a theoretical model is built of university student dropouts through expert validation, the data obtained were verified by Linear Regression with least squares ordinary. The theoretical model of attrition is validated using prediction models using Multilayer Perceptron Neural Networks and Deep Learning, through Knowledge Discover in Data Base (KDD) methodology for mining projects data. The results indicate that the technique with the highest precision rate was the backpropagation algorithm a prediction accuracy of 98.2% was obtained. I know concludes that the proposed models are adequate in terms of reliability that They are sustained under an experimental procedure. |
|---|