Abstract
In engineering activities, management of designers' knowledge is becoming crucial to share and reuse it efficiently. Toward this issue, we are studying a methodology to capture the designer's thought process, and to semi-automatically arrange it into a refined document, which addresses a particular objective with particular thought patterns. The key issues of this methodology are to acquire the thought patterns and to define relations among them. This paper reports our algorithm to acquire thought patterns from a record of designer's thought processes and any other types of documents including patent documents and academic papers. We employ a knowledge discovery approach which is recently discussed in data mining research.
