Sequential patterns analysis in a student database

This paper presents a data mining methodology to analyze the careers of students, where a career can be seen as a sequence of exams. The model is built using sequential pattern analysis and uses the algorithm SPAM. We consider an ideal career corresponding to a student which has taken each examination just after the end of the corresponding course, without delays. The frequent patterns identified by the sequential pattern analysis are then compared with the career of the ideal student. The most interesting patterns are then used to refine the analysis by using clustering techniques. Finally, we apply this methodology to a real case study and interprete the results.