Modeling is the creation of simulations in order to make judgments in advance of actual events. Modeling and simulation are used for a variety of purposes, including strategic decision-making. Specifically, user modeling is a cross-disciplinary analysis of how humans act within specific computer environments. Understanding how users will behave can assist in the building of better websites and software applications and those that work well for a wider array of users, including those with disabilities.
A primary use of user modeling is to understand how users with different attributes are likely to interact with a user interface. Attributes include skill level from novice to expert as well as other important user attributes. These include attitudes, beliefs, goals, interests, learning styles, plans, and preferences. It is possible for very different users to use the same website or software application, and user modeling is a step towards knowing what to expect. User modeling is very important in creating e-learning modules for self-tuition, since there is no teacher present to make assessments about student characteristics.
User models are created in different ways for adaptive systems and adaptable systems. An adaptable system is one over which the user is given some control. This could be through preferences or customizable elements. User modeling in an adaptable system is both performed in advance to create useful choices and ongoing as users take advantage of those choices. Data can be collected on the choices users make to guide further development. A simple example is systems with a ‘Basic’ and ‘Advanced’ interface that allows the user to choose which way to interact.
Adaptive systems are those in which the interface and/or content is structured to adapt to the user as the user’s skills, preferences, and abilities become known and change. Adaptive systems may use preferences and disabling or masking of expert components in order not to overwhelm novices and to only provide beginners with features that user modeling suggests that they need. In order to adapt, the system may rely on a variety of direct and indirect input. Direct input includes preferences, assessments, and choices. Indirect input includes user classifications that are — at least initially — based on assumptions that may be stereotypical. Examples of adaptive systems include HyperAudio®, a portable guide for museum visitors; the Office Assistant in Microsoft® Office® ’97 that popped up and offered help to users; and P-TIMS®, a financial management system.