Mining log data for the analysis of learners’ behaviour in web-based learning management systems

Nowadays, a great number of learning management systems (LMSs) both commercial and open source are widely used for educational and training purposes. Thousands of lessons are delivered through the web by utilizing some LMS and this phenomenon is expected to be escalated the following years by means of the continuous increase in demand for life long learning. In such a learning model learners, educators and educational organizations need smarter tools to face the problem of continuous feedback of the educational process in order to offer a learning environment capable to increase the learning effectiveness of the new mode of learning as well as the efficient organization of the institutional resources. Recently, the academic community uses techniques from data mining giving rise a new hot area in the e-learning field called “Educational Data Mining”. Although LMSs can hold large volumes of rich data, they provide a limited set of reporting features and they don’t support any data mining techniques. Moreover, the current data mining tools both commercial and open source are complex systems, they have not been designed for educational purposes and therefore they are difficult to be used by non-experts in this specific area. In this paper, we firstly examine the requirements for data mining facilities in LMSs. Then we describe a new approach supported by a tool for analyzing learners’ behaviour in LMSs. Finally, we describe initial results arising from the use of the tool in two undergraduate courses.