A COMPUTATIONAL MODEL FOR FORECASTING UNDERGRADUATE STUDENT PERFORMANCE BASED ON ABSENCES AND EVALUATION GRADES

One of the important issues in education is how to track students’ performance properly, taking into consideration the increasing number of students in a traditional classroom (or virtual classroom) in a class, while there is only one teacher yet. The question is how a teacher could identify students that will present some difficulty in the discipline, and understand their individual difficulties in order to help them. For such kind of identification, it is strictly necessary for the teacher has a previous knowledge about performance of each student and his absence in the class to be able to relate such past performance the impact of such past difficulties in the current discipline. In this work it is presented a model based on complex networks paradigm developed to forecast students’ performance in the course disciplines in the current term based on the grades and absences of past terms as a way to identify students that potentially will experiment some difficulty in the discipline and will present a not so good score. The initial results indicates the possibility to forecast grades of the ongoing disciplines in a term of the most students of one class. Using such grade estimative teacher may have a diagnosis of the students and would be able to give them some attention in order to help them to solve such difficulties. The data for model evaluation is the grading of undergraduate students of a computer science course.