Evaluation Method for a Human-based Evolutionary Computation System for Pattern Language Generation

Abstract

Pattern language is a method of describing knowledge to verbalize the tacit knowledge of experts and to share highly specialized knowledge with novice learners and non-specialists. By sharing specialized concepts widely in society, it is expected to support education and training of human resources in specialized fields and smooth communication in projects that cross multiple fields of expertise. Pattern language generation is usually performed synchronously by multiple people in a face-to-face manner, and the generation process is serial. Meanwhile, we previously proposed using a human-based evolutionary computation system for pattern language generation. The system allows participants to generate it asynchronously and in parallel, unlike the conventional method. However, we have not been able to show how to evaluate the output of the system. Therefore, in the paper we propose a way to evaluate a human-based evolutionary computation system for pattern language generation. The way involves design of experiment using the system.