Detección de patrones de éxito en estudios universitarios de la Universidad Continental

Abstract

The objective of the study was to detect the success patterns in the university studies of Universidad Continental students from the academic and sociodemographic information collected in the first cycle of studies. The research had a cross-sectional descriptive design, the students were from diverse professional careers that developed their students between 2012 and 2017. For the automated extraction of knowledge the methodology of data mining was used (CRISPDM) and the Clementine software, throught the multilayer perceptron neural network, it was possible to identify the variables that most impact the success and abandonment of university studies, To obtain the decision tree,algorithm C5.0 was applied and for grouping (three groups) the algorithms k-means and TwoStep, and these results were compared using a confusion matrix. The results show that the students that abandoned their university studies did not pass the fifth cycle. The variables that most influence the success of the university studies are: student`s civil status, amount of the pension, marital status of the parents and with whom the student lives in Huancayo. On the other hand, the variables that most influence the abandonment of the studies are , whit whom the student lives, the student marital status, satisfaction with the teacher’s performance and the parent`s marital status. In conclusion, the socio demographic variables are those that have the most predominance over the success and failure of university studies.