Fuzzy object-oriented user modeling: Theory, representation, and effectiveness.

摘要

User modeling aims to help users reach their goals with minimal effort and high effectiveness. User modeling is the process of gathering user goals, knowledge levels, and preferences and representing them in constructs called user models. User modeling is essential for Intelligent User Interfaces (IUI) as IUIs employ information from user modeling as the basis for interaction with users. Currently, stereotype-based user modeling is the most frequently used approach for building user modeling systems. Nevertheless, this approach has several limitations, i.e., reasoning under uncertainty, individualization, and feature identification. Therefore, to overcome the shortcomings of the conventional approach and to enhance the efficiency of user modeling in general, this dissertation extends the similarity-based fuzzy object-oriented data model to represent user models. Based on this data model, the Fuzzy Object-Oriented User Modeling (FOOUM) approach is defined. In this dissertation, the effectiveness of the FOOUM approach is examined using a simulation study. The statistical analysis of the simulation outcomes identifies that the FOOUM approach outperforms the conventional stereotype-based user modeling approach in dealing with the three limitations. Therefore, fuzzy set theory and object orientation provide sound theoretical foundations for developing a user modeling approach. However, both of these theoretical foundations contain limitations that need to be overcome in order to assure complete support for the development of a robust user modeling approach. As a result, this dissertation addresses theoretical issues and pragmatic approaches to enhance both fundamentals.