Data Mining for User Modeling

What Is User Modeling (Um)?

UM is the process of profiling users by observing them in action. It is very important to consider that the model is created by observation. It is different from asking users about how they behave when they are using a certain system. It is different from talking with expert users who may know how users are supposed to behave. In fact, users themselves usually do not know how to describe what they do, especially if they are familiar with the tasks they perform. It happens because users usually behave leaving out activities that they do not even notice they are doing. They emphasize activities that they find difficult or boring and they do not realize that there are other ordinary activities that they perform, too. It means that the true could not exactly be what they believe to be true (Hackos & Redish, 1998).

Traditionally, the different methods for capturing the behavior of a user include survey, interview, group discussion, experiment and observation and so on. However, nowadays new technologies and techniques as data mining has been introduced in order to simplify this process. Research on UM can be traced back to the early 1970s, but it was in the mid-1980 when interesting research works on this area appear. Since the early 1990s, we are witnessing and information revolution; currently, the World Wide Web and other new platforms have populated the lives of an increasing number of people with many different computing systems and large quantities of information are produced at a fast rate. These aspects tend to increase the need for user modeling and personalization.