Determination of Dropout Student Profile Based on Correspondence Analysis Technique

Abstract

The purpose of this paper is to analize how the Correspondence Analysis Techinique can be used to enhance the Exploratory Data Analysis on the drop-out for a set of educational data with a majority of the categorical type. In order to do these analisys, the independence calculation was implemented from the chi-square test, and the heatmap and perceptual map were generated with the indexes on the socioeconomic data of students. As a result of this paper is presented the relations of attraction and repulsion between socioeconomic attributes and drop-out, and the profile of the student most vulnerable to evasion is drawn. Some characteristics are: students who have already failed at least once, who do not live with their parents, elementary school in public school, low level of education of the financial responsible. The studies were carry out using a real academic database of Federal Institut of Rio Grande do Norte (IFRN) in Brazil.