Mineração de dados educacionais : predição comportamental em ambientes de educação a distância (EaD)

Abstract

The present research discusses issues such as education and educational theories to address problems encountered in Distance education institutions, with special focus on the high dropout rates of students found this type of education. Starting with this issue, it is possible, through the usage patterns of Virtual Learning Environment (VLE), identify previously problems related to evasion and propose previous diagnoses. To identify these patterns, it is taken as a tool, the Web Usage Mining and Decision Trees to predicting behavioral.This work follows the line of knowledge Discovery in Databases and uses Weka tool (free data mining software), in particular, the J48 algorithm, for this purpose. A field survey was conducted through an interview with a professional distance education to identify key signs of evasion in distance courses and the diagnoses and methods used. From the attributes identified in using the system, a prototype was designed to identify, from the records of logs recorded in the Database, prior information of the students, who perchance might exit the course. This work is aimed primarily at professionals in the field of information technology and education researchers who wish to propose technology solutions for institutions offering degrees through distance education.