Some successful applications of machine learning algorithms have produced interesting results. However, there is much open research due to the many challenges that arise for this kind of problem. First of all, data gathered at universities were not initially planned for decision making analysis and this gives rise to several different data formats or schemas which are hard to handle and process with existing mining techniques. Therefore, it is of critical importance to perform data engineering and preprocessing such that the data input is appropriate for machine learning methods. In addition, operations such as data cleaning and missing values completion are essential in order to produce accurate and meaningful models of the student’s performance.