Autonomous Learning Systems: From Data Streams to Knowledge in Real-time

14

Modelling Evolving User Behaviour with ALS

Human (including user) behaviour modelling is a challenging task and a complex phenomenon that is very important nowadays from many perspectives:

• quality of services provided;
• available technology (ubiquitous sensors, computational and wireless communication devices) makes this dream a reality;
• privacy and security;
• physics, physiology and psychology, medicine and care;
• mathematical modelling, learning, and so on.

In this chapter only the last aspect will be the focus of consideration. The ALS proposed in the book will be studied as a useful tool for modelling user behaviour in its dynamic evolution. This problem is important for computer users, users of the Internet, for the so-called assisted living directed towards the increasingly older population in the developed world and in the context of the ubiquitous computing, communication and sensory devices that are becoming an everyday reality.

14.1 User Behaviour as an Evolving Phenomenon

Traditionally, user behaviour is considered as a complex, but fixed category. The important aspect of its dynamically evolving nature is often completely ignored. In reality, the behaviour of users is a dynamic category that may change with time, it can be erratic or influenced by other factors (Iglesias et al., 2012).

A brief review of the existing literature indicates that the methods and techniques proposed and used ignore the evolving nature of the problem. For example, (Han and Veloso, 1999) ...