Universidad Autonoma Metropolitana Azcapotzalco (MEXICO)
About this paper:
Appears in: EDULEARN16 Proceedings
Publication year: 2016
Pages: 5014-5022
ISBN: 978-84-608-8860-4
ISSN: 2340-1117
doi: 10.21125/edulearn.2016.2186
Conference name: 8th International Conference on Education and New Learning Technologies
Dates: 4-6 July, 2016
Location: Barcelona, Spain
Scholar desertion is one of the aspects that most concern educational institutions, there are many studies that show many situations which can be factors that cause, in a higher or lower grade, that a student drop its studies. In Mexico, this problem results in that approximately 34% of the students drop their middle school (three years after secondary school). This work presents the use of Data Mining techniques applied to statistics of students that have or haven't concluded their middle school in Mexico. The main goal of this work is to obtain and presenting a group of models which allow to identify the main factors that cause the desertion of students at this level and based on this, take the appropriate action to prevent this situation.

Based in statistical information, Data Mining techniques are applied for generating models which allow to classify and predict the possible behavior of students compared with data of other students that ended, or didn't, their middle school and which have been previously classified according their performance, characteristics and other factors. Classification and prediction are made considering three main aspects: academic, personal and occupational, if the student had a job during its studies. Using these data, a model is obtained for each one of these characteristics. Finally, a model which gather the former aspects is developed in order to discover the situations and factors that influence the most in scholar desertion.

Statistical information contains a lot of elements which could cause that a student drops its studies, for example: socioeconomic and personal aspects, scholarship of parents and relatives, environment and social influence, activities and situations, scholar performance, school characteristics, working conditions and other.

The obtained results consist in models that allow to classify students in groups, and according their characteristics, identify or predict if they might have the risk of dropping their studies. This information can be used in order to prevent this situation and taking the necessary steps to allow them to successfully finish their studies.
Scholar desertion, educational data mining, factors for scholar desertion, preventing scholar desertion, studies dropout.