Towards establishing a score of usability evaluation

The paper deals with the problem of finding a methodology which allows obtaining a score of usability evaluation for various kinds of user interfaces. The evaluation is in principle based on a de- finition of a set of properly chosen key characteristics affecting the usability of the user interface. Such criteria are then evaluated by a group of typical users of the system. These evaluations are in principle expressed as imprecise, vague linguistic expressions stating commonly known truths representing some value of quality of use. Since the crisp numbers cannot directly express vague terms or ambiguities, theory of fuzzy logic has been chosen as the most appropriate apparatus to deal with the uncertainty. The evaluations expressed in users‘ natural language are converted to the set of normalized evaluation words. With the help of database of commonly used words, users are not limited to use their common language. Each normalized evaluation is a member of an empirical scale and can be represented by one fuzzy number whose parameters are obtained by assigning linguistic and numeric evaluation to the evaluation. A fuzzy inference system that uses expert knowledge is used to match the best possible approximation of evaluation with the knowledge included in the fuzzy rule base. The output of inference system is obtained by using common defuzzification methods. An overall usability score represents a meaningful and authentic value - a mark of the quality of particular user interface, value that can be compared to the others. Proposed methodology was successfully implemented into software application Fuzzy Usability Evaluator that has been used for evaluating usability of Web portals in Public administration.