Data Mining of Educational Data in Government Distance Learning

Abstract

Classification methods based on decision trees are used to confirm a correlation between students activity patterns in class and their final grades. By facilitating tasks like identifying participant characteristics, doing predictive performance analysis, and recognising learning kinds and patterns, Educational Data Mining (EDM) has proven to be an indispensable tool for enhancing online and distance learning (ODL). There is a significant body of research on the surroundings of universities and colleges presented in the scientific literature. However, the pedagogical paradigm used in these settings shares features with higher-level classes. In this section, we propose the application of EDM techniques for descriptive and predictive identification of interaction patterns in a governmental corporate Virtual Learning Environment (VLE), in the offer of short-term training courses in the instructional modality (with tutoring). Data were analysed regarding the interaction logs of students from two classes of a distance learning course. Classification methods based on decision trees are used to confirm a correlation between students' activity patterns in class and their final grades. Then, through clustering techniques and using the final grades as criteria, the groups of students separated according to the characteristics of interaction with the VLE and the final performance are identified. The results show that the application of EDM techniques can be used in corporate education scenarios, identifying the interaction profiles of students according to the performance obtained at the end of the course.