Modelo predictivo para la determinación del nivel de riesgo de deserción estudiantil en tecnología en electrónica de La Universidad Distrital empleando redes neuronales artificiales

Abstract

The college dropout remains a major problem in the University District, because they generate negative impacts on the social and economic sphere. Recent studies by the University District on student retention, show that the average graduation rate was 42% and the dropout rate was 58% for the period 1992-2010. This dropout rate is very high compared to the average in public universities in the country that is 39% (Vergara, 2011) (Ministry of Education, 2009). Given the magnitude of the problem of dropout, through the University District University Welfare created the Office for Student Permanence (OPEUD), who has done statistical analyzes of student dropout and survival (Quintero, 2013). The results of these studies have been taken as a reference to implement policies, as food aid, remedial courses, academic tutoring among others. However, and despite the efforts made by the University District to reduce student desertion, this remains an unresolved problem. The proposed study seeks to establish patterns to predict the level of risk of dropping out of students of the Faculty of Technology, using the technique of computational learning: Artificial Neural Networks. The data will be analyzed belong to the cohort 2008-I during the period 2008-I 2014-I. It is noteworthy that no history of studies that address the problem of dropout in the University District found using these techniques. Especially those related to artificial neural networks.