Data mining algorithms have been proved to be useful for the processing of large data sets in order to extract relevant information and knowledge. Such algorithms are also important for analyzing data collected from the users' activity users. One family of such data analysis is that of mining of log files of online applications that register the actions of online users during long periods of time. A relevant objective in this case is to study the behavior of online users and feedback the design processes of online applications to provide better usability and adaption to users' preferences. The context of this work is that of a virtual campus in which thousands of students and tutors carry out the learning and teaching activity using online applications. The information stored in log files of virtual campuses tend to be large, complex and heterogeneous in nature. Hence, their mining requires both efficient and intelligent processing and analysis of user interaction data during long-term learning activities. In this paper, we present a bi-clustering algorithm for processing large log data sets from the online daily activity of students in a real virtual campus. Our approach is useful to extract relevant knowledge about user activity such as navigation patterns, activities performed as well as to study time parameters related to such activities. The extracted information can be useful not only to students and tutors to stimulate and improve their experience when interacting with the system but also to the designers and developers of the virtual campus in order to better support the online teaching and learning.

Peer Reviewed