A model based on decision trees to predict student dropout in Private Higher Education

Abstract

The data mining techniques allow to obtain useful information that is hidden in large database that are mostly only used to perform transactional operations, as well as files that have not yet been entered into the database. The information to be exploited properly can improve decision-making and deliver competitive advantages over other companies. Due to the large amount of data with Institutions of Higher Education University in this research it is proposed to use data mining techniques to predict the desertion or abandonment in Private Higher Education. For the development of project CRIPS-DM methodology was used with commercial tool Spss v. 12.0, for which use of mining technique trees decision data were made, for which 1761 data students used Private University Cesar Vallejo, including the semesters 2009-I semester 2013-II of the professional School of Systems Engineering with 27 attributes for each that are related to the defection of the student, which were recovered from the area of academic records, and Student Affairs Computer area. For the development of the project made use of decision trees algorithm where training, validation and testing with 100 new data where an accuracy of 89% was obtained was made.