Individualizing user interfaces: Application of the Grade of Membership (GoM) model for development of fuzzy user classes

Abstract

Application of fuzzy set theory [35] provides a conceptual framework for empirical development of fuzzy user classes for measurement of computer users. Fuzzy classes generalize discrete (fixed boundary) classes by assigning scores that relate each person to each class for representing within-class heterogeneity [13, 25]. Use of fuzzy classes permits individual heterogeneity to be represented by a relatively few analytically defined types [14]. Applying the properties of fuzzy set theory to user classification will result in the definition of a user's membership within a series of fuzzy user classes within the user space. These fuzzy classes can be considered an alternative method for defining stereotypes by empirically defining potential categories into which users can be assigned. The major difference between fuzzy user classes and stereotypes lies in the application of grades of membership to directly measure simultaneous membership in multiple categories. Thus, variability can be very accurately measured and represented using fuzzy sets and grades of membership. These fuzzy classes or user types represent archetypical users or fuzzy users. Application of fuzzy set theory provides an opportunity to extend the current classification methods to measure the differences between users more accurately. This increase in accuracy assists in developing effective adaptive human computer interfaces.