Article Preview
TopUser click modeling is a relevant topic of many different fields, going from marketing to forensic, hence related work is genuinely interdisciplinary and very broad. Even within the e-learning field it is relevant for Educational Data Mining, personalized learning, tailored didactical assistance, content recommendation, classifying or clustering students, predicting scores or failure rates, and similar tasks. Especially in education, if from one side the availability of Virtual Learning Environments (VLE) and LMS has the potential to improve effectiveness of learning and reduce costs, from the other side the available data and the functionalities of popular LMS are acknowledged not to suffice for many high valued educational related analytics. For this reason, most of recent and related literature aims to enrich the available data integrating different information sources and to exploit Data Mining algorithms for developing high level tools that improve LMS user experience of learners and course designers, for example designing adaptive resources, predicting user performance or dropout, recognizing learning style, etc. Common requirement of all these tasks and of the foremost importance is using, designing and enriching feedback measurement strategies.
From a general perspective, it is possible to recognize some trends strictly related to feedback strategies in the related literature. The first research area here outlined is to expand conventional LMS with powerful analytics.
Zorrilla et al. (2008, 2010), starting from the inadequacy of the standard reporting tools in LMS and the scarcity of the feedbacks, first developed a tool called MATEP for monitoring and analyzing the behavior of users in VLE, exploiting the data in a log-based data warehouse, and then a Decision Support System (DSS) built on top of this enriched LMS to help instructors to analyze the academic progression of students .
Similarly, in Blagojević and Micić (2013) an OLAP engine with the aim of improving LMS effectiveness by predicting behavioral patterns of students and adapting accordingly the structure of the courses is proposed. The system was created based on logs from the Moodle platform.