Courseware assessment through data mining techniques

Περίληψη

 Database files and additional log files of Learning Management Systems (LMS) contain an enormous volume of data which usually remain unexploited. A new method is proposed in order to analyze these data both on the level of the courses and the learners. A new architecture based on a 3-level schema for courseware evaluation is proposed. Six measures and three metrics are used and offer useful insights into the courses themselves based on their material and usage by the learners. Two data mining techniques, classification and association rule mining, are applied to the LMS data at the first two levels. Furthermore, regression analysis is applied to the same data at the last two levels. The proposed method was successfully tested to LMS data from a Greek University. The results confirmed the validity of the approach and showed a relationship among the components of the proposed 3-level schema.