Resumen de Predicting Students’ Grade Based on Social and Content Interactions
Sheila Lucero Sánchez López, Rebeca Díaz Redondo, Ana Fernández Vilas
Predicting students’ grades is an interesting task that may be tackled using different information, like the students’ learningbackground or their previous and current performance. However, and with the widely use of the e-learning platforms, theinteraction of the students with the platforms constitutes a dynamic, huge, objective and reliable data source. Assumingthese interactions reflect the interest, feeling and way of study of each student, our proposal uses this information for theearly prediction of their performance (grades). Contrarily to other approaches, we do not use the same data analysis modelfor all the students, neither their census or learning background data. We previously define a classification scheme of fivedifferent profiles, which represent five different ways of learning and face the course. After that, a specifically created dataanalysis model is defined for each kind of student. The proposal is validated with undergraduate students of a blended-course for 3 consecutive academic years.
